SINGLE-LOCI AND MULTI-LOCI TARGETED SINGLE POINT AMPLICON

FRAGMENT SEQUENCING

CROSS REFERENCE

[0001] This application claims the benefit of U.S. Provisional Application No. 63/302,313 filed January 24. 2022, and U.S. Provisional Application No. 63/340,004 filed May 10, 2022, both of which are incorporated herein by reference in entirety.

INCORPORATION BY REFERENCE

[0002] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference. To the extent publications and patents or patent applications incorporated by reference contradict the disclosure contained in the specification, the specification is intended to supersede and/or take precedence over any such contradictoiy material.

TECHN ICAL FIELD

[0003] fhe presently disclosed subject matter relates to a high-throughput, high- resolution and low-cost method of next generation amplicon fragment sequencing of biological samples.

BACKGROUND

[0004] Liquid biopsy based on circulating cell-free DNA (cfDNA) provides a new prospect for the diagnosis, monitoring and risk assessment of a range of diseases. cfDNA molecules circulating in peripheral blood originate from dying human cells as well as from viruses, parasites, and colonizing or invasive microbes that release their nucleic acids into the blood as they die and break down (Jahr et al, 2001). Human-derived cfDNA has evolved into an indispensable biomarker in clinical practice for rapid and noninvasive diagnosis in prenatal screening, organ transplantation, and oncology (Decker and Shell, 2020; Liang et al, 2019; Sun and Yiang, 2019; Wu et al, 2020).

[0005] Although early studies did not focus on cfDNA of microbial origin (hereinafter referred to as mcfDNA), the development of circulating mcfDNA-based tests for infectious diseases has recently been gaining traction in clinical practice. An increasing number of studies have demonstrated that mcfDNA detection offers the potential to reliably identify a wide variety of infections, such as invasive fungal infection, tuberculosis, sepsis, cystic fibrosis (Rassoulian Barrett et al, 2020) and chorioamnionitis (Witt et al, 2020; for review see Man et al, 2020).

[0006] In addition to their role in infectious diseases, several studies have shown the presence of distinct cultivable bacteria in a range of cancers, including lung (Jin et al, 2019), prostate (Gorelick et al, 1988; Cohen et al, 2005), pancreas (Geller et al, 2017; Riquelme et al, 2019), and colon cancers (Bullman et al, 2019; Castellarin et al, 2012). It was only recently suggested that cancer types outside of the aerodigestive tract, such as breast (Ufbaniak et al, 2016) or brain cancer (Venkataramani et al, 2019; Zeng et al, 2019), may also harbor microbiota with distinctive compositions (for review, see Sepich-Poore et al, 2021 ), including fungi (Narunsky-Haziza et al, 2022), Both Nejman et al. (2020) and Poore et al. (2020) suggested the existence of distinct intratumoral microbiomes among >30 cancer types; these microbiomes also vary in composition at different developmental stages of the tumor, thus providing biomarkers for disease progression and prognosis for patient outcomes. As for other bacteria that are colonizing or infecting the body, the tumor associated bacteria will release distinct mclDN A in the blood stream, and this let Poore et al (2020) propose the analysis of mcfDNA from the peripheral blood as a tool to gain valuable information regarding the progression of various types of cancers.

[0007] Conventional amplicon-based sequencing approaches are routinely used to determine microbial community composition in a wide range of biological samples. The most used approach is amplicon sequencing of the 16S rRNA gene based on its variable regions, such as the V1 -V2 and V3-V4 regions (Gupta et al, 2019). Shahir et al (2020) applied 16S rRNA gene sequencing to identify region-specific composition and aerotolerance profiles of mucosally adherent bacteria in biopsy samples taken from the colon and ileum of Crohn's disease and non~IBD patients. As an alternative to 16S rRNA gene sequencing, single copy proteins encoding housekeeping genes including the genes for the DNA gyrase subunit B (gyrB) (Poirier et al, 2018), RNA polymerase subunit B (rpoB) (Vos at al, 2012; Ogier et al, 2019), the heat shock protein 60 (hsp60), the superoxide dismutase A (sodA), the TU elongation factor (tuf) (Ghebremedhin et al, 2008) and the 60 kDa chaperonin protein (cpn60) (Links et al, 2012) have been proposed as phylogenetic marker genes.

[0008] Liquid biopsy samples, especially peripheral blood, represent unique challenges for the analysis of microbial signatures. The majority of mcfDNA fragments in blood was found to be approximately 40 - 100 bp in size (Burnham et al, 2016), as was confirmed by Rassoulian Barrett et al (2020). Due to the small size of mcfDNA fragments conventional amplicon-based sequencing approaches that target DNA fragments of several hundred nucleotides (>400) are not suitable for determining the composition of colonizing or invasive microorganisms using mcfDNA from liquid biopsy samples. For example, the V1-V2 and the V3-V4 regions of the 16S rRNA gene have an average length of 437 and 443 nucleotides, respectively. Furthermore, the concentrations of plasma cfDNA in healthy individuals varies greatly, generally within the range of 0-100 ng per milliliter of plasma, sometimes exceeding 1500 ng per milliliter. Human cfDNA accounts for the vast majority (>90% or even >99%), while mcfDNA accounts for only a small fraction with 0.08%-4.85% from bacteria, ().()0%- 0.01 % from fungi, and 0.00%-0.16% from viruses/phagcs. However, it should be noted that elevated levels of mcfDNA can sometimes be observed in certain pathological conditions, including infection, sepsis, trauma, and autoimmune diseases (Han et al, 2(320). Because the analysis of mcfDNA requires deep next generation sequencing (NGS) of plasma cfDNA to overcome the limitations of small mcfDNA fragment size and low concentration, this approach is unsuitable for the testing of large patient cohorts or routine health screening.

[0009] For example, although a lot of progress has been made in reducing the cost and increasing the throughput of NGS sequencing, it remains very expensive to analyze the mcfDNA on a routine basis for community health screening and disease prognosis / diagnostics, as is routinely performed for many other health related parameters (blood cell panels, metabolic panels, etc.) or non-invasive early detection of diseases in at risk populations, such as the screening for colorectal cancer. Thus, there remains an unmet need for improved methods to accurately determine in a high throughput and cost-efficient way to detect the presence of colonizing and invasive microbes that contribute to mcfDNA present in peripheral blood as part of clinical diagnostics and community health screening. The presently disclosed subject matter provides such improved method for high resolution, high-throughput and low- cost detection of microorganisms.

SUMMARY

[0010] In one embodiment, a method is provided for amplifying microbial cell free DNA (mcfDNA). The method includes performing, on a sample comprising microbial cell- free DNA (mcfDNA), an amplification reaction using (i) one or more degenerate primers comprising complementarity to one or more conserved regions, wherein the one or more conserved regions span at least 18 nucleotides of one or more phylogenetic marker genes designated for a set of reference microbes and (ii) a second primer comprising complementarity to a repaired version of an adaptor ligated to ends of the mcfDNA, wherein at least 25 adjacent nucleotides upstream or downstream of an end of the one or more conserved regions comprise a hypervariable region, and the one or more degenerate primers are oriented to prime polymerase extension of the hypervariable region to generate amplified mcfDNA fragments.

[0011] In another embodiment, a method is provided for amplifying microbial cell free DNA (mcfDNA), that includes performing an amplification reaction on a sample comprising microbial cell-free DNA (mcfDNA) to generate amplified mcfDNA fragments using: (i) one or more degenerate primers comprising complementarity to one or more conserved regions, wherein the one or more conserved regions span at least 18 nucleotides of one or more phylogenetic marker genes designated for a set of reference microbes, and (ii) a second amplification primer comprising complementarity to an end of the mcfDNA. In some cases, at least 25 adjacent nucleotides upstream or downstream of an end of the one or more conserved regions comprise a hypervariable region, and the one or more degenerate primers are oriented to prime polymerase extension of the hypervariable region, hi some embodiments, the end of the mcfDNA can include an adaptor and the primer can include complementarity to a repaired version of the adaptor.

[0012] In some instances, the method described herein can further include sequencing the amplified mcfDN A fragments.

[0013] In some embodiments, the method can further include, rising a computer: (a) aligning the mcfDNA fragment sequences on a sequence of the one or more degenerate primers and assigning matching sequences from the hypervariable region as representative of the same microbial species; (b) for each microbial species in part (a), searching a database of the one or more phylogenetic marker genes against the mcfDNA fragment sequences and assigning the microbial species based on the closest match; and; and (c) for the one or more phylogenetic marker genes, calculating a microbial community composition based on the relative abundance of the mcfDNA fragment sequences assigned to each microbial species. In the case of multicopy phylogenetic marker genes, such as the 16S rRNA gene, the method can further include correcting for copy number variation between each species. In the case where there are two or more phylogenetic marker genes, the method can further include determining a consolidated microbial community composition by calculating a mathematical mean of the relative abundance of each species for each of the two or more phylogenetic marker genes. [0014] The methods described herein can be used to determine the presence of one or more microbial species and/or to determine a microbial community composition. In some cases, the microbial community composition comprises one or more members of Eukaryotes, bacteria, or fungi.

[0015] In other instances, a kit is provided that includes: (a) an adaptor for ligating to the ends of cfDNA; (b) one or more degenerate primers having complementarity to one or more conserved regions, and the one or more conserved regions span at least 18 nucleotides of one or more phylogenetic marker genes designated for a set of reference microbes, wherein at least 25 adjacent nucleotides upstream or downstream of an end of the one or more conserveds region comprise a hypervariable region on the one or more phylogenetic marker genes, and the degenerate primer is oriented to prime polymerase extension of the hypervariable region; (c) a primer complementary to a repaired version of the adaptor; and (d) instructions for performing an amplification reaction on mcfDNA having the adaptor-ligated ends with the one or more degenerate primers and the primer complementary to the repaired adaptor to generate amplified mcfDNA fragments. Like the methods described above, the amplified mcfDNA fragments generated in the amplification reaction using the kit can be sequenced. In addition, the mcfDNA fragments generated using the kit can be used to determine the presence of one or more microbial species and or to determine the microbial community composition according to the methods provided herein.

[0016] In the eases where the microbial community composition is calculated as described above, the method can be utilized as a screening for: tuberculosis and other diseases caused by Mycobacterium species: pulmonary infection risks and causes in cystic fibrosis patients; the risk and onset of sepsis inpatients with compromised immune systems; detection of opportunistic bacterial pathogens originating from the oral cavity that have been linked to Alzheimer's disease, pancreatic cancer and other conditions such as endocarditis; women's health issues including Chlamydia linked to mucopurulent cervicitis, pelvic inflammatory disease, tubal factor infertility, ectopic pregnancy and cervical cancer: detection and monitoring of progression in cancer; monitoring of minimal residual disease after oncology treatments; detection and monitoring of progression and minimal residual disease of breast cancer including triple negative breast cancer; detection of esophageal cancer, precancerous colonic polyps and early stage colorectal cancer, and detection and monitoring of progression and minimal residual disease of gastrointestinal cancers in general; detection and monitoring of progression and minimal residual disease in lung cancer; non-invasive analysis of the microbiome in pancreatic cancer patients to propose treatment protocols and prognostics for long-term survival; detection of Clostridium difficile infections; post-transplant bloodstream infections and Graft versus Host Disease (GvHD); detection of hospital acquired infections by emerging pathogens of clinical concern; detection of an infection in an immune compromised person; or detection of infection or inflammation of the gastrointestinal track in Irritable Bowel Disease (Crohn's disease. Ulcerative colitis); and combinations thereof.

[0017] In the methods and kits provided herein, the conserved region can have an average sequence variance score of greater than 0,175. In some cases, the hypervariable region can have an average setjnenee variance score of less than 0,075. In other instances, the hypervariable region can have an average sequence variance score of less than 0.15. In yet other cases, the hypervariable region can have an average sequence variance score of less than 0.1,

[0018] In the methods and kits, the one or more conserved regions can span 18 to 40 nucleotides, 20 to 30 nucleotides, or 22 to 28 nucleotides of the phylogenetic marker gene.

[0019] In some embodiments of the methods and kits, the at least 25 adjacent nucleotides upstream or downstream of an end of the conserved region that includes the hypervariable region is less than 150 adjacent nucleotides. The at least 25 adjacent nucleotides upstream or downstream of an end of the conserved region that includes the hypervariable region can be less than 75 adjacent nucleotides. In other embodiments, the at least 25 adjacent nucleotides upstream or downstream of an end of the conserved region that includes the hypervariable region is less than 50 adjacent nucleotides,

[0020] In the method and kit, the adaptor can be a double stranded asymmetric linker cassette comprising a 5’ asymmetrical end and a 3’ end where the two strands are complementary. The asymmetric linker cassette can be, for example, a Y-shaped linker cassete or a single arm linker cassette. In the case of the asymmetric linker cassette, the primer complementary to the adaptor is complementary to a repaired 5’ end of the asymmetric linker cassette and, in the PCR reaction, polymerase extension from the first degenerate primer results in repair of the asymmetric linker cassette,

[0021] The method can further include performing one or more reactions to repair the ends of the mcfDNA.

[0022] In the method, each of the primers in the amplification reaction can include one or more sequencing adapter sequences. In another embodiment, the method can further include adding one or more sequencing adapter sequences to the amplified mcfDNA fragments in a second PCR. or amplification reaction.

[01)23] In the methods and kits provided herein, the set of reference microbes can be eukaryotic, fungal, or bacterial, and combinations thereof. In one embodiment, the set of reference microbes are eubacterial microbes.

[0024] In the method and kit, the phylogenetic marker gene can include rpoB, epn60, 16S rRNA, or combinations thereof.

[0025] In some embodiments, the one or more degenerate primers includes primers targeting the rpoB gene, the cpn60 gene, the 16S rRNA gene, or combinations thereof.

[0026] In the method and kit, the phylogenetic marker gene can include 16S rRNA and the conserved region can include a V3, V4, or V6 region of the 16S rRNA phylogenetic marker gene.

[0027] In the methods and kits provided herein, the phylogenetic marker gene can include rpoB and the conserved region can include nucleotide positions 1327 - 1355 based on the Escherichia coli rpoB gene sequence. Alternatively, the phylogenetic marker gene can include rpoB and the conserved region includes nucleotide positions 1627 - 1652 based on the Escherichia coli rpoB gene sequence. In another embodiment, the phylogenetic marker gene includes c/w60 and the conserved region includes nucleotide positions 571-596 based on the Escherichia coli cpn60 gene sequence. In other instances, the phylogenetic marker gene includes the 16S rRNA gene and the conserved region includes nucleotide positions 785-805 based on the Escherichia coli 16S rRNA gene sequence.

[0028] In some embodiments of the method and kit, the one or more degenerate primers includes RpoBl-R.1327, RpoB6-R1630, RpoB-FI652, RpoB7-R2039, Cpn60-R57l, I6S-V4- R, or combinations thereof.

[0029] In other instances, the one or more degenerate primers includes RpoB1-R1327, Cpn60-R571, or both RpoB1-R13277 and Cpn60R571 degenerate primers.

[0030] In some embodiments of the method and kit, the set of reference microbes includes reference fungal microbes. In these instances, the method can be used to determine the presence of one or more fungi and/or to determine the fungal community composition. In this embodiment, the one or more phylogenetic marker genes comprise a human fungal phylogenetic marker gene designated for the set of reference fungal microbes, and the one or more degenerate primers comprises complementarity to a conserved region of a the human fungal phylogenetic marker gene. In some instances, the fungal phylogenetic marker gene can be nuclear ribosomal internal transcribed spacer region 1 (ITS 1 ) or nuclear ribosomal internal transcribed spacer region 2 (ITS2). The microbial community composition that can be calculated based on the percent of the sequences assigned to each species is a fungal community composition. The amplified mcfDNA fragments can include mcfDNA from one or more members of the Ascomycota, Basidiomycota and Mucoromycota, includingzl//emwt« species, Aspergillus species, Blastomyces species, Candida species, Capnodiales species, Cladospariutn species, Malassezia species, Phaeosphaeria species, Pseudozyma species, Saccharomyces species, Sporobolomyces species, Vishniacozyma species, and Yarrawia species.

[0031 ] In the methods and kits, the one or more phylogenetic marker genes can be rpoB, chaperonin protein 60 (cpn60), 16S rRNA gene, ITS1, ITS2, DNA gyrase subunit B (gtv'jg), heat shock protein 60 (Aspdfl), superoxide dismutase A protein (sod-4), TU elongation factor (tuf), DNA recombinase proteins ( including recA, recE), trrl gene that encodes for thioredoxin reductase; rim8 gene that encodes for a protein involved in the proteolytic activation of a transcriptional factor in response to alkaline pH; kre2 gene that encodes for a- .1 ,2-rnannosyItransferase; or erg6 gene that encodes for A(24)-sterol C-methyltransferase, and combinations thereof.

[0032] In one embodiment, the method or kit can further include adding in the amplification reaction a primer to determine the presence of a functional gene designated for the set of reference microbes. The functional gene primer has complementarity to a conserved region of the functional gene. In some cases, polymerase extension from the functional gene primer results in amplification of the mcfDNA only when the adaptor is ligated to a mcfDNA fragment of the mcfDNA that has the functional gene conserved region. The functional gene can be, for example, a pathogenicity factor, a PKS gene cluster essential for colibactin synthesis, or a choline trimethylaminelyase gene.

[0033] In another embodiment of the method and kit, a primer for a conserved viral gene is included in the amplification reaction, wherein the viral gene primer comprises complementarity to a conserved region of the viral gene to determine the presence of the virus. The viral gene can be a human DNA- or RNA-based oncovirus gene. The oncovirus can be one or a combination of Epstein-Barr Virus (EBV), Human Papillomavirus (I IPV), Hepatitis B virus (HBV), Human Herpesvirus-8 (HHV-8), or Merkel Cell Polyomavirus (MCPyV). In other instances, the virus is SAR.S-CoV-2 and the conserved viral gene is SARS-CoV-2spike protein. [0034] In the kit, the mcfDNA can be included in a sample. In the method and kit, the sample can be a bodily fluid, a tissue, or an extracellular bodily substance. The sample can be whole blood, a blood fraction, serum, plasma, or combinations thereof. In some instances, the sample is a biopsy sample from a solid tumor, a skin graft, a liquid biopsy samples other than blood, or combinations thereof. In one embodiment, the sample is a stool sample.

[0035] The mcfDNA can have an average fragment length of less than about 100 bp.

[0036] The percentage of the mcfDNA in the sample can be less than about 0.05%, less than about 0.1%, less than about 1%, less than about 5%, or less than about 15%.

[0037] In the cases of the method and kit where the microbial community composition is calculated, the community composition can include one or more members of Eukaiyotes, bacteria, or fungi.

[0038] 'rhe amplified mcfDNA that is generated in the methods provided herein can include mcfDNA from one or more bacterial members of: Flavobacterium sp., Staphylococcus auricularis, Pseudomonas toyotomiensis, Rheinheimera sediminis, Finegoldia magna, Parvularcula sp., Pseudomonas stutzeri, Pseudomonas soyae, Pseudomonas saponiphila, Pseudomonas sp., Peptoniphilus harei, Quisquiliibacterium sp., Azoarcus sp., Sphingopyxis terrae, uncultured Clostridiales bacterium strain UMGS460. Staphylococcus schweitzeri, Flavobacterium erciyesense, Rhodococcus yananensis, Dielzia nuxssiliensis, Cutibaclerium acnes subsp. elongalum, Anguslibacter aerolalus, Aerococcus urinae, Klebsiella quasivariicola, Comamonas fluminis, Mycobacterium tuberculosis, Mycobacterium abscessus, Mycobacterium avium, Mycobacterium chimaera, Mycobacterium leprae, Mycobacterium xenopi, Mycobacterium (para)intracellulare, Mycobacterium kansasii, Mycobacterium gilvum, Mycolicibacterium gen. nov. (‘f‘ orteitum-vaccae ” clade), Mycobacterium gen. ( “tuberculosis-simiae ’ ’ clade), Staphylococcus aureus, Staphylococcus argenteus, Staphylococcus schweitzeri, Pseudomonas aeruginosa, Burkholderia cepacia complex, Burkholderia ubonensis, Burkholderia species Abv., Burkholderia multivorans. Burkholderia pseudomultivorans, Burkholderia pseudomallei, Burkholderia mallei, Trinickia species, Burkholderia thailandensis. Haemophilus influenzae, Haemophilus parainfluenzae, Streptococcus species at the various group and species levels, Streptococcus dysgalactiae, Streptococcus pyogenes, Streptococcus mutans, Streptococcus suis, Streptococcus mitis, Streptococcus pneumoniae, Streptococcus agalactiae, Streptococcus anginosus, Streptococcus intermedius, Streptococcus constellates, Streptococcus eipii subsp. zooepidemicus, Streptococcus oralis. Streptococcus gordonii, Streptococcus uberis, Streptococcus parasanguinis, Streptococcus sanguinis Streptococcus parauberis. Streptococcus infantarius, Streptococcus Mac, Streptococcus sali varius, Streptococcus lhermophilus, Streptococcus vestibularis, Streptococcus bovis, Streptococcus gal loly liens subsp. gallofyticus, Streptococcus gallofyticus subsp. macedonicus. Streptococcus gallofyticus subsp. pasteurianus, Streptococcus equinus, Enterococcus faecalis, Enterococcus faecium, Porphyromonas gingivalis, Porphyromonas eangingivalis, Porphyromonas uenonis, Porphyromonas endodontalis, Propionibaclerium acidifaciens, Porphyromonas asaccharofytica, Porphyromonas macacae, Prevotella pallens, Prevotella histicola. Prevotella melaninogenica. Prevotella copri, Prevotella intermedia, Prevotella oral, Prevotella nanceiensis, Prevotella salivae, Prevotella nigrescens, Prevotella denticola, Prevotella buccae, Prevotella stercorea,Prevotella. oris, Prevotella disiens, Prevotella bryantii, Prevotella shahii, Tannerella/brsythia, Bacteroides fragilis, Helicobacter pylori. Chlamydia trachomatis, Neisseria meningitidis. Neisseria gonorrhoeae, Neisseria subflava. Neisseria perfla va, Neisseria jlavescens. Neisseria cinerea, Neisseria lactamica, Neisseria weaver, Neisseria zoodegmatis, Neisseria brasiliensis. Neisseria mucosa, Neisseria, animaloris, Aggregatibacter actinomycetemcomitans, Aggregatibacter aphrophilus, Aggregatibacter segnis, Saccharopolyspora species. Bacillus clausii, members of the genera Pseudoxanthomonas and Streptomyees, Fusobacterium nucleatum subsp. pofymorphum, busobacterium hwasookii, Fusobacterium canifelinum, Fusobacterium nucleatum subsp. animalis, Fusobacterium periodonlicum, Fusobacterium necrophorum subsp. funduliforme, Fusobacterium mortiferum, Fusobacterium varium, Fusobacleriuin nucleatum subsp. nucleatum, Fusobacterium ulcerans, Fusobacterium nucleatum subsp. vhwetiiil, Fusobacterium equinum.. Fusobacterium gonidiaformans, Fusobacterium necrogenes, Fusobacterium naviforme, Peptostreptococcus stomatis, Pseudonocardia asaccharofytica, Parvimonas species including Pammonas oral and Parvimonas micra, Gemella species including Gemella morbillorum, Gemeila haemolysans, Gemella palaticanis and Gemella sanguinis, Clostridium difficile, Acinetobacter baumannii. Acinetobacter lactucae, Acinetobacter pittii, Acinetobacter calcoacelicus, Acinetobacter oleivorans, Acinetobacter nosocomialis, A cinetobacter radioresistens, Acinetobacter variabilis, Acinetobacter courvalinii, Acinetobacter ursingii, Enterobacteriaceae, Escheric/iia, or Klebsiella species.

[0039] In another embodiment, a system is provided for amplifying microbial cell free DNA (mcfDDA). The system includes a reaction vessel, a reagent dispensing module, and software to execute any of the methods for amplifying microbial mcfDNA described herein, where the method is executed robotically.

[0040'j In one instance, a computer implemented method is provided for identifying a degenerate primer. The method includes using a computer and a database comprising more than one thousand DNA sequences of a phylogenetic marker gene from a set of microbes to perform the following steps: (i) identifying a highly conserved region within the DNA sequences of the phylogenetic marker gene, wherein the highly conserved region spans at least 18 nucleotides in length and has an average sequence variance score of greater than 0.175; (ii) calculating an average sequence variance score of 25-75 nucleotides upstream of the beginning of the highly conserved region and downstream of the end of the highly conserved region, wherein an average variance score of less than 0.15 is used to identify a hypervariable region; and (iii) designing a degenerate primer sequence complementary to the highly conserved DNA region based on the relative abundance of each nucleotide in the aligned phylogenetic marker gene sequences, wherein the degenerate primer sequence is oriented to prime polymerase extension of the hypervariable region. In the computer implemented method for identifying a degenerate primer, the conserved region can span 18 to 40 nucleotides, 20 to 30 nucleotides, or 22 to 28 nucleotides of the phylogenetic marker gene.

[0041] In the computer implemented method, the set of microbes can include one or more members of Proteobacteria (including representative α-, β-, y-, δ- and E-Proteobacteria), Firmicutes (including representatives for the classes Bacilli, Clostridia, Erysipelotrichia and Negativicutes), Acinetobacteria, and Fusobacteria. In another embodiment, the set of microbes can include one or more members of Ascomycota, Basidiomycota and Mucoromycota, including Altemariu species, Aspergillus species, Blastomyces species, Candida species, Capnodiales species, Cladosporium species, Malassezia species, Phaeosphaeria species, Pseudozyma species, Saccharomyces species, Sporobolomyces species, Ikshniacozyma species, and Farrow species.

[0042] In one embodiment, a degenerate oligonucleotide primer RpoB I ~R 1327 is provided consisting of a mixture of oligonucleotides having the sequences 5’ to 3’: CGRTTDCCNARRTGRTCRATRTCRTC, wherein A ~ adenine, G ~ guanidine, C ~ cytosine, T ^::: thymine, R ^:::: purine (A or G), D ^:::: not C (A, T or G), and N ^:::: any nucleotide (A, G, C or T).

[0043] In another embodiment, a degenerate oligonucleotide primer RpoB6-R1630 is provided consisting of a mixture of oligonucleotides having the sequences 5’ to 3’: TGHACRTCDCGNACY TCRWADCC, wherein A ^:::: adenine, G ™ guanidine, C ^: ™ cytosine, T thymine, R ~ purine (A or G), Y ^:::: pyrimidine (T or C), W ~ weak ( A or T), II ^:::: not G (A, T or C), D == not C (A, T or G), and N ==== any nucleotide (A, G, C or T).

[0044] In another instance, a degenerate oligonucleotide primer Cpn60-R57l is provided consisting of a mixture of oligonucleotides having the sequences 5’ to 3’: CCNYKRTCRAABYGCATNCCYTC, wherein A = adenine, G = guanidine, C = cytosine, T ^:::: thymine, R ^::: purine (A or G), Y == pyrimidine (T or C), K = amino (T or G), B ^::: not A (T, G or C), and N ~ any nucleotide (A, G, C or T).

[0045] In other embodiments, degenerate oligonucleotide primers RpoBl-RI327, RpoB6-R I630, and Cpn60-R57l are provided in which one or more of the nucleotides at primer positions represented by B, D, or N are replaced by inosine.

BRIEF DESCRIPTION OF I HI DRAWINGS

[0046] Figure 1 is a schematic of SPA fragment generation. The arrow indicates the position of the SPA primer (5 ’ to 3 *). The SPA fragment refers to the mcfDNA fragment region that will be amplified.

[0047] Figure 2 is a schematic overview of the protocol for generating single point amplification (SPA) fragments for sequencing. The various steps are numbered in order of their successive execution. Once single point amplicon fragments are generated, they are sequenced using the standard protocol for next generation paired-end Illumina sequencing.

[0048] Figure 3A is a schematic overview of the protocol for the processing of single point amplicon sequencing data for the analysis of microbial community composition. The various steps are numbered in order of their successive execution. Blastn alignment of the longest bin fragment maximizes the accuracy of microbial species identification, while readlevel normalization aims to achieve the best approximation of relative titers for microbial species identified.

[0049] Figure 3B is a schematic overview of the protocol for the processing of SPA fragment sequencing data for the analysis of microbial community composition using multiple phylogenetic identifier genes.

[0050] Figure 4 is a histogram of the lengths of the Amplicon Sequence Variants (AS Vs) resulting from SPA fragment sequencing using the RpoB6-SPA-seq-F1652 primer.

[0051] Figure 5 is a histogram of the lengths of the Amplicon Sequence Variants (ASVs) resulting from SPA fragment sequencing using the 16S-SPA-seq-V4-R primer. [0052] Figure 6 is an overview of an exemplary method used for SPA primer selection. [0053] Figure 7 A shows nucleotide statistics for the rpoB gene region 1327-1352 and degenerate sequence (GAYGAYATYGAYCAYYTNGGHAAYCG) which is the reverse complement sequence of degenerate primer RpoBl-RI 327. The relative abundance of a nucleotide at a specific position was calculated using the nucleotide sequences of 47,505 aligned unique rpoB genes from the PATR1C database and used to design the degenerate sequence, which is provided from 5’ to 3’ using the following nucleotide codes: A: adenine; G: guanidine; C: cytosine; T: thymine; R: purine (A or G); Y: pyrimidine (T or C); H: not G (A, T or C); N: any nucleotide (A, G, C or T); *: presence of an ambiguous sequence at a specific rpoB gene position. The percentages of highly conserved nucleotide sequences used to determine the consensus sequence for the degenerate primer are highlighted. The position of the region is based on the nucleotide sequence of the Escherichia coli rpoB gene.

[0054] Figure 7B shows nucleotide statistics for the epn60 gene region 571 -593 and degenerate sequence (GARGGNATGCRVTTYGAYMR.NGG) which is the reverse complement sequence of degenerate primer Cpn60-R571. The relative abundance of a nucleotide at a specific position was calculated using the nucleotide sequences of 40,989 aligned unique cpn60 genes from the PA'TRIC database and used to determine the degenerate sequence for this region, which is provided from 5’ to 3’ using the following nucleotide codes: A: adenine: G; guanidine; C: cytosine; T: thymine; R; purine (A or G); Y: pyrimidine (T or C); M: amino (A or C); V: not T (A, G or C); N: any nucleotide (A, G, C or T); *: presence of an ambiguous sequence at a specific cpn60 gene position. The percentages of highly conserved nucleotide sequences used to determine the consensus sequence for the degenerate primer are highlighted. T he position of the region is based on the nucleotide sequence of the I^cherichia coll cpn60 gene.

[0055] Figure 8 shows nucleotide statistics for the rpoB gene region 1528-1550 and degenerate sequence (CARYTNTCNCARTTYATGGAYCA). The relative abundance of a nucleotide at a specific position was calculated using the nucleotide sequences of 48, 151 aligned unique rpoB genes from the PA'TRIC database and used to design the degenerate sequence, which is provided from 5’ to 3’ using the following nucleotide codes: A: adenine; G: guanidine; C: cytosine; T: thymine; R: purine (A or G); Y: pyrimidine (T or C); N: any nucleotide (A, G, C or T); *: presence of an ambiguous sequence at a specific rpoB gene position. The percentages of highly conserved nucleotide sequences used to determine the consensus sequence for the degenerate primer are highlighted. The position of the region is based on the nucleotide sequence of the Escherichia coti rpoB gene.

[0056] Figure 9 shows nucleotide statistics for the rpoB gene region 1690-1709 and degenerate sequence (CCRATRTTNGGNCCYTCNGG). The relative abundance of a nucleotide at a specific position was calculated using the nucleotide sequences of 47,505 aligned unique rpoB genes from the PATR1C database and used to design the degenerate sequence, which is provided from 5’ to 3’ using the following nucleotide codes: A: adenine; G: guanidine; C; cytosine; T: thymine; R: purine (A or G); Y: pyrimidine (T or C); N: any nucleotide (A, G, C or T); *; presence of an ambiguous sequence at a specific rpoB gene position. The percentages of highly conserved nucleotide sequences used to determine the consensus sequence for the degenerate primer are highlighted. The position of the region is based on the nucleotide sequence of the Escherichia coli rpoB gene.

[0057] Figure I 0A is a graph showing the variance of the 75 bp region located upstream (5’) of region recognized by the RpoB 1-R.l 327 primer sequence. The variance score is calculated as the variance of the percentage of the nucleotide adenine, guanidine, cytosine and thymine at each position of the rpoB gene, calculated for the 47,505 rpoB genes which aligned on the RpoB I -R 1327 primer. A lower number is indicative tor more variance, while a higher number is indicative for less variance and a more conserved DNA sequence. The maximum theoretical variance score, plotted on the Y-axes, is 0.25 ( 100% conserved nucleotide at a position). The region recognized by the RpoB I -R1327 primer (nucleotide numbers 76 - 101 on the X-axes) is indicated by the arrow.

[0058] Figure 10B is a graph showing the variance of the 75 bp region located downstream (3’) of region recognized by the RpoBl-F1352 primer sequence. The position of the region recognized by the RpoB I -Fl 352 primer (nucleotide numbers 1 - 26 on the X-axes) is indicated by the arrow.

[0059] Figure 11 is a graph showing the number of unique SPA fragments with length of 25, 50, 75, 100 and 200 nucleotides for the regions located upstream or downstream of the annealing site for the RpoBl -R 1327 and RpoBl-F1352 primer, respectively.

[0060] Figure 12 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Mycobacterium tuberculosis, Mycobacterium tuberculosis subsp. africanum, Mycobacterium canetti i and Mycobacterium orygis strains identified by the presence of SPA fragments My I and My2. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB 1 -R 1327 primer annealing site. [0061 ] Figure 13 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Mycobacterium avium strains identified by the presence of SPA fragments My8 and My9. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl-RI 327 primer annealing site.

[0062] Figure 14 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Mycobacterium strains identified by the presence of SPA fragments My 17 and My 18, The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site,

[0063] Figure 15 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Staphylococcus strains identified by the presence of SPA fragments Sa1,, Sa2, Sa3 and Sa4. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl-R 1327 primer annealing site.

[0064] Figure 16 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Pseudomonas strains identified by the presence of SPA fragments Pa1, Pa2, and Pa4. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl-R1327 primer annealing site.

[0065] Figure 17 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Burkholderia pseudomallei group strains identified by the presence of SPA fragments Bpml, Bpm2, Bpm3 and Bed, The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl-R1327 primer annealing site,

[0066] Figure 18 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Haemophilus influenzae and Haemophilus para influenzae strains identified by the presence of SPA fragments Hi I , H2, Hi6 and Hi7, The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB 1 -R1327 primer annealing site.

[0067] Figure 19 is a schematic showing the whole genome-based Average Nucleotide Identity ( Arahal, 2014) between representative Streptococcus dysgalactiae and Streptococcus pyogenes strains identified by the presence of SPA fragments St2, St3 and St4. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site.

[0068] Figure 20 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Streptococcus gordonii. Streptococcus oligofermentans, Streptococcus mitis and Streptococcus oralis strains identified by their SPA fragments. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site.

[0069] Figure 21 is a schematic showing the 'whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Streptococcus anginosus, Streptococcus constellatus and Streptococcus intermedins strains identified by the presence of SPA fragments St 14 to St 17. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site.

[0070] Figure 22 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Streptococcus lhermophilus, Streptococcus vestibularis and Streptococcus salivarius strains identified by the presence of SPA fragments St30, St31 and St32. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site.

[0071] Figure 23 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Streptococcus gallolyticus subsp. gallolyticus, Streptococcus gallolyticus subsp. Macedonians, Streptococcus gallolyticus subsp. pasteurianus and Streptococcus equinus strains identified by the presence of SPA fragments St33, St34 and St35. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site.

[0072] Figure 24 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Enterococcus faecalis and Enterococcus faecium strains identified by the presence of SPA fragments Efl, E12, Ef3 and Ef4. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB 1-R1327 primer annealing site.

[0073] Figure 25 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Porphyromonas strains identified by the presence of SPA fragments Pg.1 to Pg9. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 R ppriomBe1r-R an1n3e2a7ling site.

[0074] Figure 26 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Bacteroides fragilis strains and related species identified by the presence of SPA fragments Bfl , Bf2 and Bf3. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site. [0075] Figure 27 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Helicobacter pylori strains identified by the presence of SPA fragments I -Ip 1 , Hp2 and Hp3. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl-Rl 327 primer annealing site.

[0076] Figure 28 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Aggregatibacter strains identified by the presence of unique SPA fragments. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl-Rl 327 primer annealing site.

[0077] Figure 29 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Acinetobacter baumannii strains and related species identified by the presence of their unique SPA fragments. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327

[0078] Figure 30 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Acinetobacter baumannii strains and related species identified by the presence of their unique SP.A fragments. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB I -R 1327 primer annealing site. SPA fragment ‘ref indicates a reference strain included.

[0079] Figure 31 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Acinetobacter baumannii strains and related species identified by the presence of their unique SPA fragments. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site. SPA fragment ‘ref indicates a reference strain included.

[0080] Figure 32 is a schematic showing the whole genome-based Average Nucleotide Identity (Arahal, 2014) between representative Klebsiella and related strains which share SPA fragment Ent2 (see Table 38). The 50 nucleotide SPA fragments upstream of the RpoB6- R1630 priming site are identified as SPA fragment “Ent” with a numerical identifier and with an asterisk symbol indicating that the SPA fragment w ^r as generated from the region upstream of the RpoBl-Rl 630 priming site. SPA fragment ‘ref indicates a reference strain included.

[0081 ] Figure 33A is a phylogenetic tree of Escherichia coll and related species based on the sequences of 50 nucleotide SPA fragments generated from the region upstream of the RpoB1-R1327 priming site. Clusters of Escherichia coli phylotype B2 sand D strains are indicated.

[0082] Figure 3313 is a phylogenetic tree of Escherichia coli and related species based on the sequences of 50 nucleotide SPA fragments generated from the region upstream of the RpoB6-R1630 priming site. Clusters of Escherichia coli phylotype B2 sand D strains are indicated.

[0083] Figure 33C is a phylogenetic tree of Escherichia coli and related species based on the combination of 50 nucleotide SPA fragments sequences generated from the regions upstream of the RpoB1-R1327 and RpoB6~R1630 priming sites. Clusters of Escherichia coli phylotype B2 sand D strains are indicated,

[0084] Figure 34 A is a schematic showing the whole genome-based Average Nucleotide Identity (ANI) comparison for the Eaecalibacteritan species present in the consortium.

[0085] Figure 3413 is a schematic showing the whole genome-based Average Nucleotide Identity (ANI) comparison for the Bacteroides ovatus strains present in the consortium.

[0086] Figure 35 is a graph showing the simulation of mcfDNA fragment length distribution. Average fragment lengths of 40. 60, 80 and 100 base pairs were used in the simulations, respectively. For each simulation, the size distribution of a million mcfDNA fragments around a truncated normal distribution was used.

DETAILED DESCRIPTION

[0087] The presently disclosed subject matter now will be described more fully hereinafter. The presently disclosed subject matter may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. Indeed, many modifications and other embodiments of the presently disclosed subject matter set forth herein will come to mind to one skilled in the art to which the presently disclosed subject matter pertains having the benefit of the teachings presented in the descriptions provided herein. Therefore, it is to be understood that the presently disclosed subject matter is not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. [0088] Following long-standing patent law convention, the terms “a,” “an?" and “the” refer to “one or more” when used in this application, including the claims. Thus, for example, reference to “a sample” includes a plurality of samples, un less the context clearly is to the contrary, and so forth.

[0089] Throughout this specification and the claims, the terms “comprise,” “comprises,” and “comprising” are used in a non-exclusive sense, except where the context requires otherwise. Likewise, the terms “having” and “including” and their grammatical variants are intended to be non-limiting, such that recitation of items in a list is not to the exclusion of other like items that can be substituted or added to the listed items.

[0090] For the purposes of this specification and claims, the term “about” when used in connection with one or more numbers or numerical ranges, should be understood to refer to all such numbers, including all numbers in a range and modifies that range by extending the boundaries above and below the numerical values set forth. The recitation of numerical ranges by endpoints includes all numbers, e.g., whole integers, including fractions thereof, subsumed within that range (for example, the recitation of 1 to 5 includes I . 2, 3, 4, and 5, as well as fractions thereof] e.g., 1.5, 2.25, 3.75, 4.1, and the like) and any range within that range. In addition, as used herein, the term ’’about", when referring to a value can encompass variations of, in some embodiments +/-20%, in some embodiments +.<-10%, in some embodiments +/- 5%, in some embodiments +/-!%, in some embodiments +/-0.5%, and in some embodiments +/-0.1%, from the specified amount, as such variations are appropriate in the disclosed compositions and methods. Alternatively, particularly with respect to biological systems or processes, the term can mean within an order of magnitude, preferably within 5-fold, and more preferably within 2-fold, of a value. Where particular values are described in the application and claims, unless otherwise stated the term “about” meaning within an acceptable error range for the particular value should be assumed.

[0091] Throughout this specification and the claims, the term “subject” includes humans and animals and can be used interchangeably with the term “human” and the term “patient”.

[0092] The terms “SPA fragment” and “SPA fragment sequence” are herein used interchangeably.

[0093] The terms “PCR reaction” and “amplification reaction” are herein used interchangeably. [0094] The term “phylogenetic marker gene” as used herein means any conserved gene from any organism, including but not limited to bacteria, fungi, parasites, and viruses, that is suitable for phylogenetic identification.

[0095] There are a wide range of diseases where microbial community analysis, especially of the gut microbiome, provides important information regarding the disease or its treatment options. This includes conditions such as IBD (Ananthakrishnan et al, 2017), metabolic diseases (Boulange et al, 2016), diseases of the central nervous system (Bhattacharjee and Lukiw, 2013 ) and cancer where the interaction of the gut microbiome can provide clues regarding the response to specific treatments including immune checkpoint inhibitors (Gopalakrishnan et al, 2020; Sepich-Poore et al, 2021). Deep microbial metagenome sequencing is the most informative approach when it comes to microbial community analysis, as it will provide detailed information regarding community composition as well as the key functions encoded by the community members. Unfortunately, despite major breakthroughs in metagenome sequencing technologies to reduce its costs, it is currently still too expensive lor routine screening purposes of human associated microbial communities in large population screenings. Another disadvantage of deep microbial metagenome sequencing is the need for relatively large amounts of high-quality microbial DNA. This has hindered its application to study the microbial communities associated with liquid and solid biopsy samples, where only a small fraction of the total DNA is of microbial origin.

[0096] The amplification and subsequent sequencing of phylogenetic marker genes provides an alternative, cheaper high throughput method for microbial community analysis. For example, in tissue biopsy samples where there is sufficient concentration of DNA having average fragment length of about 5,000 bp or more, amplification-based sequencing approaches have been successfully applied to identify differences in microbial communities between healthy individuals and patients suffering from a wide range of diseases. Advantages of the amplification and subsequent sequencing method include that it requires significantly less DNA than metagenome sequencing, and because specific DN A primers are used to amplify phylogenetic target genes, there is little contamination with host DNA, making this method suitable to analyze the microbial communities associated with tissue biopsy samples, from which small amounts of high molecular weight DNA can be obtained. However, analysis of microbial signatures in liquid biopsy samples, especially peripheral blood samples, results in additional challenges as compared to tissue biopsy samples, due to the low concentration of mciDNA having small fragment sizes. [0097] For example, in plasma, human cfDNA accounts for the vast majority of cfDN A (>90% or even >99%), while mcfDNA accounts for only a small fraction with 0.08%-4.85% from bacteria, 0,00%-0,0.l% from fungi, and 0.00%-0.16% from viruses/phages (Han et al, 2020). However, the percentage of mcfDNA compared to cfDNA should be placed in the context of the human genome size and the size of an average microbial genome, with sizes of 6.4 billion and approximately 6 million nucleotides, respectively, therefore providing similar coverage. Thus, mcfDNA represents an important signal that is largely being ignored in liquid biopsy testing.

[0098] The intrinsic properties of cfDNA. and mcfDNA, especially its small fragment sizes, make its analysis for disease detection and monitoring chal lenging. More than 70% of plasma cfDNA is smaller than 300 bp, with an average size of 170 bp (Fernandez-Carballo et al, 2019). However, the size of mcfDNA fragments was found to be significantly smaller, approximately 40-100 bp (Burnham et al, 2016), as was confirmed by Rassoulian Barrett et al (2020). As a result of this size limitation, conventional amplicon-based sequencing approaches including 16S rRNA gene and rpoB gene amplicon sequencing that target DNA fragments of several hundred nucleotides, are not suitable for determining the composition of colonizing or invasive microorganisms using mcfDNA from peripheral blood and other liquid biopsy samples. The small size of mcfDNA makes it nearly impossible to use mcfDNA in ampliconbased sequencing protocols, such as 16S rRNA gene sequencing, leaving no other option than high-cost and low-throughput NGS sequencing.

[0099] To overcome the above-mentioned limitations, the present inventors developed a single point amplification sequencing approach that exploits the combination of a degenerate primer for a conserved region of a marker gene located adjacent to a phylogenetic hypervariable region of the gene for a wide range of microbes. The method is based on the targeted amplification of high-resolution phylogenetic identifier fragments from mcfDNA, which comprises a fraction of the total cfDNA isolated from, for example, biopsy samples. To generate the phylogenetic identifier fragments, a hypervariable DNA region with high phylogenetic resolution is targeted. The hypervariable region located next to the highly conserved region that functions as a primer annealing site as is illustrated in Figure I . In the methods disclosed herein, the fragments resulting from specific amplification of the hypervariable DNA regions are referred to as SPA fragments.

[00100] In various embodiments, methods and kits are provided herein for generating the SPA fragments. The methods and kits provided herein can be used to determine the presence of one or more microbial species and/or to determine one or more microbial community compositions. In the methods and kits provided herein, the set of reference microbes can be eukaryotic, fungal, or bacterial, and combinations thereof. In one embodiment, the set of reference microbes are eubacterial microbes.

[00101 ] In the methods of the invention, the length of the SPA fragment is determined by the distance between the end of the mcfDNA fragment and the 3 ’-end of the primer annealing site. Only mcfDN A fragments that contain the primer annealing site will give SPA fragments, which can be subsequently sequenced and used for high resolution phylogenetic identification and analysis of community composition.

[00102] In one aspect of the invention, the degenerate primer is used in combination with an adaptor, such as, for example, an asymmetric linker cassette which is attached to the 3’ ends of all the cfDNA fragments in the sample. A PCR amplification reaction is performed using the degenerate primer and a primer complementary to the 5’ asymmetrical end of the linker cassette. The degenerate primer is designed to allow for DNA synthesis into the hypervariable region. However, successful PCR amplification of the hypervariable region occurs only when the asymmetric linker cassete is repaired. In a PCR reaction, the asymmetric linker cassette will be repaired only when located downstream from the degenerate primer annealing site, i.e, when the asymmetric linker cassette has been ligated to a mcfDNA fragment that contains the conserved region of the phylogenetic marker gene. In this manner, microbial DNA fragments that originate from the hypervariable region are selectively amplified.

[00103] In one example of the invention, to overcome the above-mentioned limitations for determining microbial profiles from mcfDNA, such as, for example, mcfDNA in liquid biopsy samples, the present inventors developed a unique approach that exploits the phylogenetic resolution of a hypervariable region of the rpoB gene,. In another example of the invention, the present inventors developed a unique approach that exploits the phylogenetic resolution of V3-V4 hypervariable region of the 16S rRNA gene. In contrast to commonly used amplicon sequencing, in which regions between two conserved DN A sequences are targeted for PCR amplification, the methods provided herein use a single conserved DNA sequence as the primer annealing site to initiate PCR amplification. The amplification initiated from this single conserved DNA sequence allows for targeted amplification of the hypervariable region located adjacent to the primer annealing site, independent of the size of the fragment, followed by sequencing of the amplified fragment. In another example of the invention, the phylogenetic resolution of a hypervariable region of the chaperonin cpn60 gene is used in the presently disclosed methods. This method may be referred to herein as Single Point Amplification (SPA) fragment sequencing.

[00104] Alternative embodiments of the invention include use of a conserved DNA sequence as the primer annealing site for more than one site on a phylogenetic marker gene or for a site on two or more different phylogenetic marker genes in a single amplification reaction. In one instance, two degenerate primers targeting different regions of the rpoB gene are included in the presently disclosed methods. In another instance, a degenerate primer for both the cpn60 and the rpoB gene are included in the presently disclosed methods. The use of two or more degenerate primers for annealing to two or more conserved regions on a single or two different phylogenetic marker genes may be referred to herein as “multi-loci SPA fragment sequencing”.

[00105] In the specific examples provided herein the RNA polymerase subunit B (rpoB) gene and the chaperonin 60 (cpn60) gene were used, but it should be noted that the SPA fragment sequencing method is very broadly applicable to conserved housekeeping genes, including, but not limited to, the prokaryotic genes coding for the DNA gyrase subunit B (gyrB), the heat shock protein 60 (hsp60), the superoxide dismutase A protein (sodA), the TU elongation factor (tuf), and the DN/\ recombinase proteins (including recA, recE). The SPA fragment sequencing method can also be applied on the Prokaryotic 16S rRNA gene, for instance to amplify (part of) the V1-V2 or V3-V4 hypervariable region. The SPA fragment sequencing method can also be applied on the Eukaryotic internal transcribes spacer (ITS) regions ITS1, which is located between the LBS and 5.8S rRNA genes, and ITS2, which is located between the 5.8S and 28S rRNA genes. The SPA fragment sequencing method can also be applied to genes that are unique to pathogenic fungi including the trrl gene that encodes for thioredoxin reductase: the rim8 gene that encodes for a protein involved in the proteolytic activation of a transcriptional factor in response to alkaline pH; the kre2 gene that encodes for a- 1,2-mannosy I transferase; and the erg6 gene that encodes for A(24)-sterol C - methyltransfera.se (Abadio et al, 2011); or any conserved gene from any organism, including bacteria, fungi, parasites, and viruses that is suitable for phylogenetic identification. This includes conserved genes from the human DNA-based oncoviruses, more specifically the Epstein-Barr Virus (EBV), Human Papillomavirus (HPV), Hepatitis B virus (IIBV), Human Herpesvirus-8 (HHV-8), and Merkel Cell Polyomavirus (MCPyV) (Mui et al, 2017). One or a combination of any conserved housekeeping gene can be used in the presently disclosed methods. [00106] Advantages of the disclosed SPA fragment sequencing method include an increase in the diversity of hypervariable regions that can be targeted for amplicon analysis as the method only requires one neighboring conserved region to bind the primer (compared with the two required by dual primer approaches). As such, the SPA fragment sequencing method is more adaptable, flexible, and offers greatly improved resolution over current methods. In addition, the multi-loci SPA sequencing methods include the advantage of improving phylogenetic resolution for the identification of the community members on the species and subspecies level, as is highlighted in EXAMPLE 13. Further, the multi-loci SPA sequencing methods provide an internal control for improved error correction in the SPA fragment amplification and sequencing process, as similar results for community species abundances are expected independent of the phylogenetic identifier gene.

[00107] In addition to the degenerate primer for the conserved region, an adaptor such as, for example, an asymmetric linker cassette, can be used to introduce a DNA sequence that is targeted by a second primer in the PCR amplification reaction. In one embodiment, to avoid amplification of any DNA fragment flanked by two adaptors, the adaptors are “defective” or in other words “asymmetric”. This can be accomplished by designing an adaptor as an asymmetric linker cassette where the strand that serves as the template for primer annealing is missing. Typical asymmetric linker cassette configurations include, but are not limited to:

1. A “Y”-shaped linker cassette, where two single stranded DNA fragments that are only partially complementary are annealed. This results in an asymmetric linker cassette where one end is double stranded, allowing for ligation, but where the other end is comprised of two single stranded non-complementary DNA strands.

2. A “single arm” linker cassette, where a shorter single stranded DNA fragment is annealed to the complementary 3 ’-end of a longer single stranded DNA fragment. This results in an asymmetric linker cassette with a single stranded the 5 ’-end and a double stranded 3’ -end.

[00108] Ideally, the single strands of the asymmetric linker cassette are complementary over a stretch of about at least 16 nucleotides with an annealing temperature of approximately 50®C or higher, allowing for a linker cassette that is stable at room temperature. The single strand of the asymmetric linker can also contain 6 random nucleotides that constitute a Unique Molecular Identifier (UM I) to correct PCR induced errors and improve sequencing accuracy. To avoid sell-ligation, in one example, the asymmetric linker cassette includes a 3 'sticky end. The 3'sticky end can be formed by a single nucleotide, such as, for example. thymine. To avoid undesirable repair of the asymmetric linker cassette initiated from the shorter single stranded DNA fragment, the terminal 3’ nucleotide can be a dideoxy nucleotide that functions as a chain-elongating inhibitor of DNA polymerase.

[00109] In a PGR reaction, the asymmetric linker cassette will only be repaired when located downstream from the degenerate primer annealing site. For purposed of the specification and c l aims, the term "repaired" when used in the con text of the asymme tric linker cassette, means that a new DNA strand is created in the PCR reaction that is complementary at the 5' end of the asymmetric linker cassette. DNA synthesis initiated from the degenerate primer into the asymmetric linker cassette will restore the defective DNA strand complementary to the S’-end of the linker and in this manner the asymmetric linker cassette is repaired. In subsequent PCR cycles this strand is used for primer annealing, allowing for the amplification of the hypervariable region. To allow for sample multiplexing and sequencing, the resulting amplicons can be further amplified in a second PCR reaction to introduce two Unique Dual Indexes (UDI), one at each end of the amplicons, and, for example, the Illumina sequencing anchors P5 and P7.

[00.110] In one embodiment of the invention, the method includes one or more of the following steps as detailed in Figure 2:

1. Isolation of cfDNA using standard protocols. Cell-free DNA can be extracted from 0.5 ml. blood plasma using the typically yielding 0. 1 ng to 10 ng to be used for sequencing. cfDNA can also be isolated from urine, saliva, stool and other biopsy samples,

2. End repair and 5’-phosphorylation of cfDNA fragments followed by the 3’ addition of a deoxy-adenine to create a 3’-sticky end formed by a single adenine nucleotide using standard protocols. A typical protocol to process cfDNA includes end repair (blunting and 5' phosphorylation), 3' A-taiiing, followed by adaptor ligation. The fragment ends are repaired by blunting and 5' phosphorylation with a mixture of enzymes, such as T4 polynucleotide kinase (PNK) and T4 DN A polymerase (T4 DNA pol). This end repair step is followed by 3' A-tailing at 37 ^1> C using a mesophilic polymerase such as Klenow Fragment 3'-5’ exonuclease minus (Head et al, 2014). Many commercial kits are available to perform this step.

3. Ligation of the adaptor, which in this case is an asymmetric linker cassette, using T4 DNA ligase. Many commercial kits are available to perform this step, including the NEB NEBNext® Ultra™ II Ligation kit or the IDT xGen™ DNA Lib Prep MC kit. I’he sequences of the oligonucleotides used for the design of the asymmetric linker cassette, referred to as SPA-casl and SPA-cas2, are provided in Table 1 along with other primer sequences that can be utilized in the methods and kits provided herein. Annealing of these two partially complementary single stranded DNA fragments results in a ⁱ 'Y”-shaped or single arm DNA linker cassette. On one end, the two strands of the linker cassette are not complementary. On other end, where the two strands are complementary, the linker cassette includes a 3'sticky end formed by a single thymine nucleotide. Due to the sticky ends, the only possible ligation is between cfDNA fragments and asymmetric linker cassettes, while self-ligation of linker cassettes and repaired cfDNA fragments is blocked.

4. Single point linker cassette repair, PCR is performed on the ligation product using the following primers: (a) the SPA I -amp primer that recognizes the repaired 5’ asymmetrical end of the linker cassette; (b) one or more primers that recognize the primer annealing site specific for the conserved region of the one or more phylogenetic marker genes. DNA amplification initiated from the gene-specific SPA primer will result in the repair of the asymmetric linker cassette but only when this cassete is bound to a cfDNA fragment that contains the primer annealing site on the conserved region. This will be limited to mciDNA fragments that contain the targeted region of the phylogenetic marker gene such as, for example, positions 1630 - 1 ,652of the rpoB gene, which is absent in Eukaryotic DNA including human DNA. As such no human cfDNA fragments flanked by asymmetric linker cassettes will be repaired. In EXAMPLE 1 herein below, the RpoB6-SPA-seq-F1652 primer, which recognizes the rpoB gene sequence between positions 1630 - 1652, and the 16S-SPA-seq-V4-R primer which recognizes the 16S rRNA gene sequence between positions 785-805, were validated. In addition, due to the direction of the amplification reaction from the gene specific primer such as, for example, the forward RpoB6-SPA-seq-F1652 primer, only linker cassettes that are bound to the region upstream of the targeted region of the phylogenetic marker gene will be repaired (c.g., position 1630 - 1652 of the rpoB gene).

5. Once the asymmetric linker cassete has been repaired, the primer (SPAl-seq- F primer) that recognizes the repaired 5’ asymmetrical end of the linker cassette can anneal and PCR amplification is initiated. In the case of the RpoB6-SPA-seq-F1652 primer this will result in the amplification of DNA sequences located downstream of position 1652 of the rpoB gene. The forward (SPA1-seq-F) and reverse (e.g. RpoB6- SPA-seq-F1652) primers include a 5’ extension corresponding to the Illumina Read-1 and Read-2 sequences, respectively, to allow sequencing library preparation.

After the amplification step has been completed, an optional enrichment step can be performed by annealing a 5’-biofinilated version of the one or more gene specific primers (e.g., RpoB6-SPA-seq-F1652 primer) followed by capturing the hybridized primer on magnetic streptavidin beads. Subsequently, the non-captured DNA fragments are washed away, and the targeted DNA fragments are eluted using a NaOH solution. After neutralization and precipitation, these fragments are ready for the construction of sequencing libraries.

As an alternative to the affinity-based enrichment step, an enrichment PCR protocol can be used to reduce background amplification of human DNA fragments resulting front nonspecific primer annealing. The enrichment PCR uses the SPA-amp primer in combination with one primer annealing to the conserved region extended by a tew nucleotides (e.g. RpoB6-F1649) compared to the primer used in STEP 4 (e.g. RpoB6- SPA-seq-F1652). Neither primer used in the first step of the enrichment PCR contains the Illumina Read-1/2 extension.

6. In a second PCR reaction (PCR2), Unique Dual Indexes (UDI) and Illumina sequencing anchors (P5 and P7) are added to the amplified SPA fragments using P5- 15-Rdl and P7-l7-Rd2 primers (see Table 1 ). The PCR2 is performed using unique sets of UDI for each sample, subsequently allowing the pooling of the libraries, after which fragments are paired-end sequenced using NGS Illumina sequencing, e.g, on the Illumina NEXTSEQ 1000 (Illumina, Inc.. San Diego, CA). This approach will result in sequenced fragments that all share the sequence of the gene specific primer (e.g., RpoB6-SPA-seq-F1652 primer) followed by sequences that vary in length and nucleotide composition. Sequences derived from the same microorganisms will be identical except for the length of the sequenced fragment, which will vary as a function of the distance between the gene specific primer annealing site (e.g., RpoB6-SPA-seq- F1652 primer) and the end of the mcfDNA fragment.

Table 1: Overview of primer sequences. The following nucleotide codes were used: A: adenine; G: guanidine; C: cytosine; T: thymine; R: purine (A or G); Y; pyrimidine (T or C); W: weak (A or T); S: strong (G or C); M: amino (A or C); K: keto: (G or T); B: not A (T, C. or G); II: not G (A. T or C); D: not C (A, T or G); N: any nucleotide ( A, G, C or T). The extended primer sequences used for multiplex Illumina sequencing are shown in to/rcs. _* indicates a phosphorothioated DNA base to protect the linker from 3’ end degradation.

[00111 ] in some embodiments of the invention, the processing and analysis of the SPA fragment sequences includes one or more of the following steps as shown in Figure 3 A:

1. Reads are filtered based on read quality. Error correction is done using software such as DADA2 (Callahan et al, 2016), which makes use of a parametric error model. The remaining error-corrected reads of different lengths are deduplicated while recording the number of duplicates by sequence for calculating community composition.

2. Unique SPA fragments are aligned on the sequence of the RpoB6-SPA-seq-F1652 primer forming bins of matching sequences representative for the same species.

3. The database of bacterial rpoB genes is searched for the longest read in each bin of matching sequences for species identification. If a fragment does not match exactly to the database of bacterial rpoB genes, the closest match species is assigned, noting the likelihood of a false match.

4. Community composition is calculated based on the percent of reads assigned to each bin, taking into consideration the number of duplicate reads identi fied in step

[001 12] Additional primers besides those derived from the RpoB6-FI 652 and the I6S-V4-R primers can be used for SPA fragment sequencing. EXAMPLE 2 describes the design of alternative rpoB gene specific primers. A RpoB 1-R1327 primer, which recognizes the rpoB gene sequence between positions 1327 - 1352 (positions based on the Escherichia coli rpoB gene sequence) and allows for generation of SPA fragments upstream of this region, was validated in silico for the phylogenetic resolution of the sequences of 50 nucleotide Single Point Amplification (SPA) fragments as described in EXAMPLES 3 to 9. In EXAMPLE 7 a RpoB6-R1630 primer, which recognizes the tpoB gene sequence between positions 1630 - 1652 and allows for generation of SPA fragments upstream of this region, was validated, and EXAMPLE 10 describes the combined use of the RpoBl~R1327 primer and RpoB6-R1630 primer for improved identification of members of the Enterobacteriaceae. EXAMPLE 13 describes the Cpn60-R571 primer, which recognizes the cpn60 gene sequence between position 571-593, (position numbers based on the Escherichia coli cpn60 gene sequence). In another embodiment of the invention, a method is provided for multi loci SPA fragment sequencing. Use of two or more different gene-specific SPA primers in the same amplification reaction such as, for example, the RpoB1-R1327 and Cpn60-R571 primers is detailed in EXAMPLE 14. One example of a protocol for the method of amplifying mcfDNA provided herein is generally illustrated in Figure 2 and is as follows:

1. Isolation ofclDNA using standard protocols.

2. End repair and 5 ’-phosphorylation of cfDNA fragments followed by the 3’ addition of a deoxy-adenine to create a 3’-sticky end formed by a single adenine nucleotide using standard protocols.

3. Ligation of an adaptor, which in this embodiment is an asymmetric linker cassette created by annealing the primers SPA-casl and SPA-cas2, using T4 DNA ligase.

4. Single point linker cassete repair. To generate multi loci SPA fragments, multiplexing PCR is performed on the ligation product using three primers: (a) the SPA 1 -amp primer that recognizes the repaired 5’ asymmetrical end of the linker cassette; (b) a primer that recognizes the primer annealing site specific for the conserved region of the first phylogenetic marker gene, such as the RpoB6~FI 652 primer; and (c) a primer that recognizes the primer annealing site specific for the conserved region of the second phylogenetic marker gene, such as the 16S-V4-R primer. Alternatively, the RpoBl-1327R primer, the Cpn60-R571 primer, or combinations of these primers can be used. These primer sequences are provided in Table 1.

5. Once the asymmetric linker cassette has been repaired, the primer (SPA 1 -amp primer) that recognizes the repaired 5’ asymmetrical end of the linker cassette can anneal and PCR amplification is initiated. In the case of the reverse RpoB6-F1652 and Cpn60- R571 primers, this will result in the amplification of DNA sequences located downstream of position 1652 of the rpoB gene and upstream of position 571 of the cpn60 gene, respectively. An enrichment PCR. protocol can be used to reduce background amplification of human DNA fragments resulting from nonspecific primer annealing.

6. In a follow up PCR step, adapter sequences are added to the amplified SPA fragments using the primers RpoB 1 -SPA- seq-Rl 327, Cpn60-SPA-seq-R571 and SPAl-seq-F (see Table 1 ). In a second PCR reaction (PCR2), UDI and sequencing anchors are added to the amplified SPA fragments using the primers P5-I5-Rd1 and P7- I7-Rd2 (see T able 1 ), The PCR2 is performed using unique sets of UDI for each sample, subsequently allowing the pooling of the libraries, after which fragments are paired- end sequenced using NGS Illumina sequencing, e.g, on the Illumina NextSeq 1000 (Illumina, Inc., San Diego, CA). Phis approach will result in sequenced fragments that share the sequence of either the RpoB6-SPA-seq-F 1652primer or the Cpn60-SPA-seq- R571 primer, followed by sequences that vary in length and nucleotide composition. Sequences derived from the same microorganisms and extended from the same primer will be identical except for the length of the sequenced fragment, which will vary as a function of the distance between the respective primer annealing site and the end of the mcfDNA fragment.

[00.113] In one instance, the processing and analysis of the SPA fragment sequences includes the following steps:

1. Similar to single loci SPA fragment sequencing, the reads are filtered based on read quality. Error correction can be done using software such as DADA2 (Callahan et al, 2016), which makes use of a parametric error model. The remaining error-corrected reads of different lengths can be deduplicated while recording the number of duplicates by sequence for calculating community composition,

2. Multi loci SPA fragment sequencing can include a step to deconvolute the reads on the phylogenetic gene level. Unique SPA fragments are aligned on the sequences of the RpoB1-R1327 primer or the Cpn60-R571 primer and sorted in gene specific “buckets”. This is schematically shown in Step 1 of Figure 3B. Subsequently, the sequences of each bucket are sorted into birrs of matching sequences representative for the same species. In a next step, the rpoB and cpn6() gene databases are searched for the longest read in each bin of matching sequences for species identification. If a fragment does not match exactly to the database entries, the closest match species is assigned, noting the Likelihood of a false match.

3. For each phylogenetic gene, the community composition is calculated based on the percent of reads assigned to each species, taking into consideration the number of duplicate reads identified in step 1 .

[001 14] To reconcile the outcomes obtained for the SPA fragments obtained 'from different phylogenetic identifier genes, their results are compared and consolidated into a consensus community description (species and their relative abundances), as is schematically shown in Step 2 of Figure 3B.

[001 15] hi one embodiment of the invention, the reconciliation process of Step 2 in Figure 3B works as follows:

1. To phylogenetically identify the community members, SPA fragments that provide the highest level of phylogenetic resolution are prioritized. Thus, SPA fragments that allow for species level identification have priority over SPA fragments that allow for identification at the genus level. For example, a subset of SPA fragments from gene 1 and gene 2 both specifically identify species A, confirming its presence as a community member. However, a second subset of SPA fragments from gene I identifies the closely related species B and D, while a second subset of SPA fragments from gene 2 is specific at the species level and indicates that only species B is present. It is therefore concluded that species B is present. Similar, a third subset of SPA fragments from gene 1 identifies the presence of speci es C , while a thi rd subset of SPA fragments from gene 2 identifies the presence of the closely related species C, species E and species F. Therefore, it is concluded that species C is present.

2. To determine the abundances of the community members, the mean of the relative abundance for each species (as determined using the SPA fragments from each of the different phylogenetic identifier genes) is calculated.

[00116] The utility of the methods of the invention is exemplified in EXAMPLES I - 14 of the present disclosure. For example, in EXAMPLE 1 of the present disclosure, the inventors demonstrate that the primers RpoB6-SPA-seq-Fl 652 and 16S-SPA-seq-V4-R can be used to generate unique SPA fragments from tnefDNA present in blood that allowed for bacterial identification on the species level based on homology to the rpoB gene and the 16S rRNA gene, respectively. In EXAMPLE 2 of the present disclosure, the inventors demonstrate that a 50 nucleotide length cutoff enabled in silico generation of 20,919 unique SPA fragments covering the rpoB gene region upstream of the RpoBl-R1327 primer annealing site. The generated SPA fragments provided sufficient phylogenetic resolution to enable identification of many bacteria at the species level. These 50 nucleotide SPA fragments were generated from 50,569 unique rpoB gene sequences present in the PATRIC database (Wattam et al, 2014). Increasing this length to 75 nucleotides had only a marginal effect on the phylogenetic resolution of this method (22,603 unique fragments). The 50 nucleotide fragment size was selected based on the average length (40-100 nucleotides) of mcfDNA fragments. It should be noted that larger fragments will also be generated for each species, further improving the resolution for the phylogenetic identification.

[001 17] EXAMPLES 3 to 9 demonstrate that, despite their relatively short size, the sequences of the 50 nucleotide long SP A fragments covering the rpoB gene region upstream of the RpoB 1 -Rl 327 primer annealing site allow for high resolution phylogenetic identification at the bacterial species level of many clinical ly relevant bacterial isolates.

[00118] EXAMPLE 10 describes a simulation showing that mcfDNA fragments with an average length of 60 base pairs can be reliably used to identify strains present at 0.5% or above in a known gut microbial community at the species and subspecies level. The species and subspecies are detectable in liquid biopsy samples, including peripheral blood. On average, strain abundances measured based on SPA fragments were within 1 .4% of the actual abundance. For strains with less than 1% abundance, the average error was 1 .8%, ranging from 0.1% to 7.2%.; for strains with an abundance of 1% or higher, the average error was 1.2%, ranging from <0.1% to 4.5%.

[00119] EXAMPLE 11 describes an experiment to determine the phylogenetic accuracy of the SPA fragments generated using the RpoB I -R 1327 primer in EXAMPLE 10. The results shows that the SPA fragments have very high phylogenetic specificity to reliably classify bacteria at both the taxonomic genus and species level.

[00120] EXAMPLE 12 is an experiment designed to access how the sensitivity and specificity of the SPA fragment sequencing methods compare to the current method of deep metagenome sequencing of cfDNA fragments followed by taxonomic classification using readbased nietagenome analysis methods. The simulations described in EXAMPLE 12 using deep metagenome sequencing of cfDNA fragments followed by taxonomic classification of mcfDNA using read-based metagenome analysis methods show that current read-based tools are unsuitable for taxonomic classification of the short sequencing reads obtained from mcfDNA. As such the current approach lacks the sensitivity and specificity to provide meaningful insights for disease detection and progression monitoring. Overcoming this limitation would require very deep sequencing and assembly of short reads into larger fragments. In addition to higher sequencing costs, limitations in the assembly of short sequencing reads render the current approach unsuitable for scalable application to the routine analysis of microbial patterns in biopsy samples.

[00121 ] EXAMPLE 13 describes identification of a degenerate primer comprising complementarity to a conserved region spanning position 571 to 593 of the cpntiO gene (position numbers based on the Fscherichla coli cpn60 gene, “Cpn60-R571 primer”) for SPA fragment sequencing. The results described in EXAMPLE 13 show that the simulated community compositions using rpoB gene-derived SPA fragments and cpn60 gene-derived SPA fragments are very similar. In addition, and unexpectedly, it was discovered that the Cpn60-R571 primer can be used in combination with the RpoB 1-R 1327 primer in the SPA fragment sequencing methods of the present disclosure to improve the phylogenetic resolution based solely on the rpoB gene. Based on this result a new method is provided, referred to as multi loci SPA fragment sequencing, which combines SPA fragments from multiple phylogenetic identifier genes to analyze the composition of microbial communities. The results of EXA MPLE 13 show that the multi loci SPA fragment sequencing method using two or more phylogenetic identifier genes, such as the rpoB and cpn60 genes, can have advantages over the SPA fragment sequencing method using a single locus. Such advantages include: (1) provision of an internal sample control for the SPA fragment amplification and sequencing process, as similar results for community species abundances are expected independent of the phylogenetic identifier gene; and (2) improvement in phylogenetic resolution for the identification of the community members on the species and subspecies level, as was highlighted in EXAMPLE 13.

[00122] The clinically relevant bacterial isolates that can be identified using the methods of the invention include, but are not limited to, blavobacterium sp., Staphylococcus auricularis, Pseudomonas toyotomiensis, Rheinheimera sediminis, Finegoldia magna, Parvularcula sp., Pseudomonas stutzeri, Pseudomonas soyae, Pseudomonas saponiphila, Pseudomonas sp., Peptoniphilus harei, Quisquiliibacterium sp., Asaamts sp., Sphingopyxis terrae, uncultured Clostridiales bacterium strain UMGS460, Staphylococcus schweitzeri. Flavobaclerium erciyesense, Rhodococcus yananensis, Dielzia massi liens is, Cutibaclerium acnes subsp. elongatum, Angustibacter aerolatus, Aerococcus urinae, Klebsiella quasivariicola, Comamonas fluminis, Mycobacterium tuberculosis, Mycobacterium abscessus, Mycobacterium avium, Mycobacterium chimaera, Mycobacterium leprae, Mycobacterium xenopi. Mycobacterium (paraflntracellulare, Mycobacterium kansasii, Mycobacterium gilvum, Mycolicibacterium gen. nov. ( “fortuitum-vaccae ” clade), Mycobacterium gen. (“iuberculosis-simiae ” clade), Staphylococcus aureus, Staphylococcus argenteus, Staphylococcus schweitzeri, Pseudomonas aeruginosa, Burkholderia cepacia complex, Burkholderia ubonensis, Burkholderia species Nov., Burkholderia multivorans, Burkholderia pseudomullivorans. Burkholderia pseudomallei, Burkholderia mallei, Trinickia species, Burkholderia thailandensis, Haemophilus influenzae, Haemophilus parainfluenzae. Streptococcus species at the various group and species levels. Streptococcus dysgalactiae, Streptococcus pyogenes. Streptococcus mutans, Streptococcus suis, Streptococcus mitis, Streptococcus pneumoniae, Streptococcus agalactiae. Streptococcus anginosus, Streptococcus intermedius. Streptococcus constellatus. Streptococcus equi subsp. zooepidemicus. Streptococcus oralis, Streptococcus gordonii, Streptococcus uberis, Streptococcus parasanguinis. Streptococcus sanguinis Streptococcus parauberis, Streptococcus infantarius, Streptococcus iniae, Streptococcus salivarius, Streptococcus thermophilus, Streptococcus vestibularis, Sireptococcus bovis. Streptococcus gallolyticus subsp. gallolyticus, Streptococcus gallolyticus subsp. macedonicus, Streptococcus gallolyticus subsp. pasteurianus, Streptococcus equinus. Enterococcus faecalis, Enterococcus /aecium, Porphyromonas gingivalis, Porphyromonas eangingivalis, Porphyromonas uenonis, Porphyromonas endodontalis, Propionibacterium acidijaeiens, Porphyromonas asaccharofylica, Porphyromonas macacae, Prevotella pa/'lens; Prevotella histieola, Prevotella melaninogenica, Prevotella copri, Prevotella intermedia, Prevotella oral, Prevotella nanceiensis, Prevotella saiivae, Prevotella nigrescens, Prevotella dentieola, Prevotella bueeae, Prevotella stereorea, Prevotella oris. Prevotella disiens, Prevotella bryanlii, Prevotella shahii, Tannerellajbrsythia, Bacteroides fragilis, Helicobacter pylori, Chlamydia trachomatis, Neisseria meningitidis. Neisseria gonorrhoeae, Neisseria subflavq. Neisseria perflava, Neisseria flavescens. Neisseria cinerea, Neisseria lactamica, Neisseria weaver, Neisseria zoodegmatis, Neisseria brasiliensis, Neisseria mucosa, Neisseria animaloris, Aggregatibacter actinomycetemcomitans, Aggregatibacter aphrophilus, Aggregatibacter segnis, Saccharopolyspora species, Bacil/m clausii, members of the genera Pseudoxanthomonas and Streptomyces, Fusobacterium nucleatum subsp. poiymorphum, Fusobacterium hwasookii, Fusobacterium canijelinum, Fusobacterium nucleatum subsp. animalis, Fusobacterium periodonticum, Fusobacterium necrophorum subsp. funduliforme, Fusobacterium mortiferum, Fusobacterium varium, Fu.sobacteriiim nucleatum subsp. nucleatum, Fusobacterium ulcerans, Fusobacterium nucleatum subsp. vincentii. Fusobacterium equinum, Fusobacterium gonidiaformans, Fusobacterium necrogenes, Fusobacterium naviforme, Peptostreptococcus stomatis, Pseudonocardia asaccharolytica, Parvimonas species including Parvimonas oral and Parvimonas micra, Gemella species including Gemella morbillorum, Gemella haemolysans, Gemella palaticams and Gemella sanguinis, Clostridium difficile, Acinetobacter baumannii, Acinetobacter lactucae, Acinetobacter pitlii, Acinetobacter calcoaceticus, Acinetobacter oleivorans, Acinetobacter nosocomialis, .4 cinetobacter radioresistens, Acinetobacter variabilis, Acinetobacter courvalinii, Acinetobacter ursingii, and members of the Enterobacteriaceae, including Escherichia and Klebsiella species.

[00123] This phylogenetic identification of many clinically relevant bacterial isolates at the species level represents a significant improvement over methods such as Kaiju (Menzel et al, 2016) or Kraken (Wood and Salzberg, 2014), which are being used for sequence-read based identification of microorganisms represented by the mcfDNA at the genus level. As is well documented for many pathogenic bacteria, including zWjwfeac/mum species, optimal patient treatment protocols including the use of antibiotics are species-level specific, showing the importance of the level of phylogenetic resolution that is uniquely obtained with the single point amplicon sequencing approach provided herein. Furthermore, by targeting genes that are absent or sufficiently different from the host genome, such as genes conserved in pathogenic fungi that are absent from the human genome (Abadio et al, 201 1), the method provided herein can also be used to detect the presence of Eukaryotic infections, such as those caused by parasitic fungi and amoeba. Candidate fungal genes for SPA fragment sequencing include: trrl that encodes for thioredoxin reductase; rimS that encodes for a protein involved in the proteolytic activation of a transcriptional factor in response to alkaline pH; kre2 that encodes for a- 1,2-mannosy'ltransferase; and erg6 that encodes for A(24)-sterol C-methyltransferase (Abadio et al, 2011).

[00124] In certain instances, disease phenotypes caused by bacteria will depend on the presence of virulence/'pathogenicity factors located on mobile genetic elements, including conjugative and/or mobi le plasmids, phages, and pathogenicity islands that can be horizontally transferred between bacteria, as is the case for Escherichia coli, Salmonella, Klebsiella. Listeria, Bacillus, pyogenic streptococci and Clostridium perfringens, among others (for review, see Gyles and Boerlin, 2014). As the result of horizontal gene transfer, in some instances phylogenetic information on species composition will be insufficient to predict disease pathology, and therefore needs to be complemented with information on community functionality. SPA fragment sequencing provides the flexibility to address both phylogenetic identification and community functionality: by selecting a degenerate primer that recognizes a conserved DNA region of a specific function, the same protocol outlined in Figure 2 and Figures 3A and 3B is broadly applicable for SPA amplification and sequencing of functional genes.

[00125] For instance, the presence in Esc/ierichia coll of the PKS pathogenicity island encoding, among other virulence factors, for genotoxic colibactin synthesis has been linked to increased risk for developing colorectal cancer (Pleguezuelos-Manzano et al, 2020). By designing a primer for SPA fragment amplification that specifically targets the PKS gene cluster essential for colibactin synthesis, the presence of genotoxic Escherichia coli strains (Pleguezuelos-Manzano et al, 2020) can be determined and combined with phylogenetic information for risk assessment of colorectal cancer.

[00126] Pan-cancer analyses recently revealed caneer-type-specific fungal ecologies and bacteriome interactions (Narunsky-Haziza ei al, 2022). By designing a primer for SPA fragment amplification that specifically targets a human fungal phylogenetic marker such as the nuclear ribosomal internal transcribed spacer region 1 (ITS!) or region 2 (ITS2), the presence of human pathogenic fungi can be determined and combined with bacterial phylogenetic information to for risk assessment of cancer. The amplified mcfDNA that can be generated in the methods provided herein can include mcfDNA from fungal species including one or more members of the Ascomycota, Basidiomycota and Mwcorowycota, including Altemaria species, Aspergillus species, .Blastomyces species, Candida species, Capnodiales species, Cladosporium species, Md/assezza species, Phaeosphaeria species, Pseudozyma species, Saccharomyces species, Sporobolamyces species, Pishniacozyma species, and Yarrowia species. The methods for amplifying mcfDNA provided herein can also be used for detecting viral DNA. For example, a primer for a conserved viral gene can be included in the amplification reaction, where the viral gene primer includes complementarity to a conserved region of the viral gene to determine the presence of the vims. The viral gene can be a human DNA- or RNA-based oncovirus gene. Assessing the risk and better understanding the cause of cancer can be improved by designing primers for SPA fragment amplification that specifically target conserved genes present in human oncoviruses. For example, the method can be used for determining the presence of human DNA-based oncoviruses such as, but not limited to, the Epstein-Barr Vims (EBV), Human Papillomavirus (HPV), Hepatitis B vims (HBV), Human Herpesvirus-8 (HHV-8), and Merkel Cell Polyomavirus (MCPyV).

[00127] In one aspect of the invention , phylogenetic and functional information can be obtained simultaneously by including both one or more degenerate primers that target the phylogenetic identifier gene(s) and a primer that targets a functional gene in the same reaction for the SPA fragment amplification step (Figure 2. step 4). This approach may be referred to herein as multiplex SPA for the simultaneous detection of multiple targets in a single reaction. Thus, the method for amplifying mcfDNA provided herein can further include in the amplification reaction a primer for a functional gene designated for the set of reference microbes, wherein the functional gene primer comprises complementarity to a conserved region of the functional gene, to determine the presence of the functional gene. The functional gene can be, but is not limited to, a pathogenicity factor, a PKS gene cluster essential for colibactin synthesis, or a choline trimethylaminelyase gene.

[00128] Since 100,000 sequencing reads represent the standard depths for ampliconbased sequencing for complex microbial community analysis, the latest Illumina NEXTSEQ instruments allow for an unprecedented number of samples to be sequenced in parallel. For example, the Illumina NEXTSEQ 6000 allows to theoretically collect 20 billion reads with a single run, which would correspond to 100,000 paired-end sequenced samples.

[00129] In addition to monitoring of specific diseases, SPA fragment sequencing can be useful as part of the general health screening. Unlike the stool microbiome, the microbiome of colonizing and infecting bacteria will be relatively stable, with changes occurring when the relation between host and microbes is changing. This includes situations of new invasions by infectious and colonizing microorganisms, such as the formation of stomach ulcers, the formation of intestinal polyps/ adenomas and their progression into malignancies, gastrointestinal diseases including Irritable Bowel Disease (IBD), various tumors and their specific microbiomes including pancreatic cancer, lung cancer and cervical cancer, Central Nervous System (CNS) diseases including multiple sclerosis (MS) and Alzheimer's disease, minimal residual disease (MRD) monitoring, and other diseases characterized by dysbiottc and inflammatory microbiomes such as cystic fibrosis or tuberculosis, and general risk monitoring of infections in patient populations with a compromised immune system, positioning SPA fragment sequencing as an ideal tool for risk monitoring, early detection, prognostics and evaluation of disease progression. Contrary to PCR. based detection methods that monitor for the presence of specific bacteria, SPA fragment sequencing provides an “open” diagnostics approach to detect any bacterium or fungus based on the presence of its mcfDNA in peripheral blood. Figures 4 and 5 show the distribution of SPA fragment lengths generated using primers targeting the rpoB gene and the 16S rRNA gene, respectively,

[00130] In one aspect of the invention, SPA fragment sequencing can provide an important non-invasive method for (early) detection and identification of infectious and colonizing bacteria using mcfDNA from peripheral blood samples, which can subsequently be linked to a broad range of diseases, including: screening for tuberculosis and other diseases caused by Mycobacterium species; determining pulmonary infection risks and causes in cystic fibrosis patients; determining the risk and onset of sepsis in patients with compromised immune systems; detection of opportunistic bacterial pathogens originating from the oral cavity that have been linked to Alzheimer's disease, pancreatic cancer and other serious conditions such as endocarditis; women's health issues including Chlamydia linked to mucopurulent cervicitis, pelvic inflammatory disease, tubal factor infertility, ectopic pregnancy and cervical cancer; detection and monitoring of progression of cancer; monitoring of minimal residual disease after oncology treatments; detection and monitoring of progression and minimal residual disease of breast, cancer including triple negative breast cancer, detection of esophageal cancer, precancerous colonic polyps and early stage colorectal cancer, and detection and monitoring of progression and minimal residual disease of gastrointestinal cancers in general; detection and monitoring of progression and minimal residual disease in lung cancer; non-invasive analysis of the microbiome in pancreatic cancer patients to propose treatment protocols and prognostics for long-term survival; detection of Clostridium difficile infections; post- transplant bloodstream infections and Graft versus Host Disease (GvHD); detection of hospital acquired infections by emerging pathogens of clinical concent; detection of an in fection in an immune compromised person; or detection of infection or inflammation of the gastrointestinal track in Irritable Bowel Disease (Crohn's disease. Ulcerative colitis); and combinations thereof. Therefore, SPA fragment sequencing represents a quantum leap forward to apply mcfDNA sequencing as a high-resolution, high-throughput and low-cost routine test in disease detection, patient monitoring, risk assessment and large-scale population screenings using mcfDNA informed biomarkers. For example the microbial footprint obtained with SPA fragment sequencing combined with the mutational footprint and methylation footprint that are currently being used as biomarkers for the detection, monitoring and prognostics of cancers, will provide a powerful tool for improved early detection and monitoring of progression of various types of cancer. It is expected that including the microbial footprint will increase the specificity'' and selectivity of screening tests, e.g. for the detection of early stage adenomas and carcinomas in colorectal cancer. Furthermore, once unique SPA fragments have been identified that correlate with the detection of specific diseases and monitoring of their progression, their sequences can be used to develop species-specific PCR-based screening assays as part of diagnostic platforms.

[00131] In addition to using mcfDNA from blood, the SPA fragment sequencing approach provided herein is applicable to analyze microbial DNA compositions in any sample type, especially when in samples having low amounts of small fragment microbial DNA. This includes biopsy samples from solid tumors, skin grafts, and other liquid biopsy samples besides peripheral blood, as well as mcfDNA present in stool samples.

[00132] In other instances, the methods and kits provided herein can be used for SPA fragment sequencing as a non-invasive method for (early) detection and identification of infectious and colonizing fungal microbes using mcfDNA from biological samples as described herein. For example, the set of reference microbes in this case includes reference fungal microbes. The method can be used to determine the presence of one or more fungi and/or to determine the fungal community composition. The one or more degenerate primers included in the amplification reaction in this embodiment includes complementarity to a conserved region of a human pathogenic fungal gene or DN A region designated for the set of reference fungal microbes. The conserved human pathogenic fungal gene or DNA region is herein referred to interchangeably for the purposes of the specification and claims as a "fungal phylogenetic marker gene”. In some instances, the fungal phylogenetic marker gene can be ITS! or ITS2. The microbial community composition that can be calculated based on the percent of the sequences assigned to each species is a ftmgal community composition. The amplified mcfDNA fragments can include mcfDNA from one or more members of the Ascomycota, Basidiomycota and Mucoromycota, including Alternaria species, Aspergillus species, Blastomyces species, Candida species, Capnodiales species, Cladosporium species, Malassezia species, Phaeosphaeria species, Pseudozyma species, Saccharomyces species, Sporobolomyces species, Pishniacozyma species, and Yarrowia species.

[00133] In the SPA fragment sequencing method, a DNA region is identified in a suitable phylogenetic marker gene that has the following characteristics:

1. Presence of a highly conserved DNA region to design a degenerate primer for annealing to the phylogenetic marker gene. 2. Adjacent to the primer annealing site the presence of a highly variable DNA region with high phylogenetic resolution. This region will become part of the single point amplicon (SPA) fragment.

[00134] An overview of an exemplary SPA primer design method is shown in Figure 6. For each phylogenetic marker gene, such as rpoB, cpn60, 16S rRNA, ITS1. ITS2, gyrB, tuf or other phylogenetic marker gene or conserved housekeeping gene including, but not limited to, those used by CheckM (Parks ei al, 2015), 50-100 species are initially selected that cover the prokaryotic diversity, including members of the phylum Proteobacteria (including representative α-, β~, y-, δ- and ε-Proteobacteria), the phylum Firmicutes (including representatives for the classes Bacilli, Clostridia, Erysipelotrichia and Ncgativi cutes), and the phyla Acinetobacteria and Fusobacteria. Marker genes for these species are aligned using a multiple sequence alignment tool like ClustalW. The SPA algorithm is subsequently used to identify conserved regions as putative annealing sites for primer candidates by looking for the highest “average sequence variance” scores over 25 nucleotide-long DNA regions among this limited set of sequences. This is performed as follows:

• Determine the percent of nucleotides for each nucleotide (GATC) at each position.

• Calculate the variance of the percentages at each position.

• Calculate region variance as the average of the variances of each position in the region.

[00135] A completely conserved nucleotide position will have 100% of one nucleotide and 0% for the other three nucleotides, and a variance of 0.25. A completely non-eonserved region will have 25% of each nucleotide and a variance of 0. Primer candidates are prioritized based on their “average sequence variance” scores.

[001.36] Primer candidates are evaluated for key properties including the level of primer degeneracy and annealing temperature (>50°C). The sequences from the complete curated marker gene database are aligned to these conserved regions to determine their nucleotide compositions. The conservation of their 3’ nucleotide (must be >99% conserved among entries) and their “average sequence variance” scores are calculated (highly conserved regions have the highest score) and used to rank, and select primer leads, prioritizing primers with the highest score.

[00137] In the next step, using a curated marker gene database, an algorithm (referred to as “SPA algorithm” in Figure 6) is used to determine the “average sequence variance” for the regions adjacent to the primer annealing site. Primers with adjacent 25 nucleotide-long and 50 nucleotide-long regions with ideally an average sequence variance of <0.15 and <0.075, respectively, are prioritized based on the lowest score. The algorithm also identifies the resolution of phylogenetic identification for the regions adjacent to each primer lead by determining the number of unique SPA fragments. SPA primers with the highest phylogenetic resolution are added to the SPA primer repository.

[00138] Figure 7A shows nucleotide statistics for the rpoB gene region 1327-1352 and degenerate sequence (GAYGAYATYGAYCAYYTNGGHAAYCG) which is the reverse complement sequence of degenerate primer RpoB.l-R1327. In this specific example, the relative abundance of at nucleotide at a specific position was calculated using the nucleotide sequences of 47,505 aligned unique rpoB genes from the PATR1C database and used to design the degenerate sequence, which is provided from 5 ’ to 3 ’ using the following nucleotide codes: A: adenine; G: guanidine; C: cytosine; T: thymine; R: purine (A or G); Y: pyrimidine (T or C); H: not G ( A, T or C): N: any nucleotide (A, G, C or T); *: presence of an ambiguous sequence at a specific tpoB gene position. The percentages of highly conserved nucleotide sequences used to determine the consensus sequence for the degenerate primer are highlighted. The position of the region is based on the nucleotide sequence of the Escherichia coll rpoB gene.

[00139] Figure 7B shows nucleotide statistics for the cpn60 gene region 571 -593 and degenerate sequence (GARGGNATGCRVTTYGAYMRNCKi) which is the reverse complement sequence of degenerate primer Cpn60-R517. The relative abundance of a nucleotide at a specific position was calculated using the nucleotide sequences of 40,989 aligned unique cpn60 genes from the PATR1C database and used to determine the degenerate sequence for this region, which is provided from 5’ to 3’ using the following nucleotide codes: A: adenine; G: guanidine; C: cytosine; T: thymine; R: purine (A or G); Y: pyrimidine (T or C); M: amino (A or C); V: not T (A, G or C); N: any nucleotide (A, G, C or T); *: presence of an ambiguous sequence at a specific cpn60 gene position. The percentages of highly conserved nucleotide sequences used to determine the consensus sequence for the degenerate primer are high lighted. The position of the region is based on the nucleotide sequence of the Escherichia coli cpn60 gene.

[00140] In the next step, the proposed degenerate primer sequences are matched to the human genome sequence and the number of hits with increased number of allowed mismatches is determined. To minimize annealing to human genomic DNA, a primer should ideally have two or more mismatches with the human genome. [00141] Various modifications and variations of the disclosed methods, compositions, and uses of the invention will be apparent to the skilled person without departing from the scope and spirit of the invention. Although the invention has been disclosed in connection with specific preferred aspects or embodiments, the invention as claimed should not be unduly limited to such specific aspects or embodiments.

[00142] The present invention may be implemented using hardware, software. or a combination thereof and may be implemented in one or more computer systems or other processing systems. In one aspect, the invention is directed toward one or more computer systems capable of carrying out the functionality described herein.

EXAMPLES

[00143] The following Examples have been included to provide guidance to one of ordinary skill in the art for practicing representative embodiments of the presently disclosed subject mater. In light of the present disclosure and the general level of skill in the art, those of skill can appreciate that the following Examples are intended to be exemplary only and that numerous changes, modifications, and alterations can be employed without departing from the scope of the presently disclosed subject matter.

EXAMPLE 1

SPA fragment sequencing using the 16S rRNA gene and the rpo# gene as phylogenetic markers

[00144] As representative examples, the SPA sequencing approach was successfully demonstrated for the rpoB gene and the 16S rRNA gene as an example of a single-copy and multi-copy phylogenetic marker, respectively.

[00145] To validate the RpoB6-F1652 primer and the 16S-V4-R primer for SPA fragment amplification from the rpoB gene and the 16S rRNA gene, the following protocol was followed. Following the steps outlined in Figure 2, cfDNA isolation was performed using the Qiagen QIAamp ccfDNA/RNA Kit on 1 ,0 ml blood plasma from healthy volunteers,

[00146] To confirm the presence of mcfDNA in the blood samples, total cfDNA was isolated on 1.0 ml blood plasma and deep sequencing was used to determine the percentage of mcfDNA. In the case of these healthy donors, the percentage of mcfDN A was approximately 0.5% of the total cfDNA (data not shown). This is considerably lower than typically found in blood samples from e.g. cancer patients, where this ranged between approximately 1% to 4% (Poore et al, 2020).

[00147] Subsequently, following the supplier's instructions the xGen™ DNA I..ib Prep MC kit (IDT) was used for end repair plus 5 '-phosphorylation on 10 ng cfDNA fragments followed by the 3’ addition of a deoxy-adenine to create a 3’~sticky end of a single adenine nucleotide (Step 2), after which 20 ng of the asymmetric SPA-linkcr-lJ Ml-Y was ligated to the repaired cfDNA fragments (Step 3) in a total volume of 16 pl.

[00148] The sequences of the two single stranded DNA fragments, SPA-casl and SPA-cas2, was used to create the asymmetric SPA-linker-UMI-Y linker cassette are listed in Table 1. The linker cassette was created by the following procedure. First, by annealing equal amounts (4 nmol) of SPA-casl and SPA-cas2. The mixture is first heated for 2 min. at 95 °C, then for I 0 min, at 65°C, 10 min, at 37°C, and finally 20 min. at room temperature. The mixture is kept on ice or stored at 4°C.

[00149] To repair the asymmetric linker cassette, a PCR reaction, referred to as PCR 1 , was performed on the ligation product using two primers: (a) the SPAl-seq-F primer that recognizes the repaired 5’ asymmetrical end of the linker cassette; (b) a primer that recognizes the primer annealing site specific for the conserved region of the phylogenetic marker gene, in this example the RpoB6-SPA-seq~Fl652 primer. The forward (SPAl-seq-F) and reverse (e.g. RpoB6-SPA-seq-F1652) primers include a 5’ extension corresponding to the Illumina Read-1 and Read-2 sequences, respectively, to allow sequencing library preparation. The PCR.1 was performed in 25 pl reaction containing lx KAPA FliFi HotStart ReadyMix, 0.2 pM of each primer, and the Linker-cfDNA ligation products. The reaction was run in a thermocycler using the following program: 1 cycle at 95°C for 10 min, 10 cycles at 98°C for 20 sec, 65°C to 50°C for 30 sec and 72°C for 15 sec, 35 cycles at 98°C for 20 sec, 60°C to 50°C for 30 sec and 72°C for 15 sec, and 1 cycle at 72°C for 1 min. A similar protocol was followed for creating SPA fragments from the 16S rRNA gene using the 16S-seq-V4-R primer.

[00150] Once the asymmetric linker cassette was repaired, the SPAl-seq-F primer that recognizes the repaired 5' asymmetrical end of the linker cassette can anneal and PCR.1 amplification is initiated. In the case of the RpoB6-SPA-seq-F1652 primer this will result in the amplification of DNA sequences located downstream of position 1352 of the rpoB gene.

[00151] In a second PCR reaction (PCR2), Unique Dual Indexes (UDI) and Illumina sequencing anchors (P5 and P7) were added to the amplified SPA fragments using P5-I5-Rdl and P7-I7-Rd2 primers (see Table I). The PCR2 was performed in 25 pl reaction containing lx KAPA 1 lil’i HotStart ReadyMix, 0.2 pM of each primer, and PCR1 bead cleaned products. The reaction was run in a thermocycler using the following program: 1 cycle at 95°C for 3 min, 8 cycles at 95°C for 30 sec, 55°C for 30 sec and 72°C for 30 sec, and 1 cycle at 72°C for 5 min.The PCR2 was performed using unique sets of UDI for each sample, subsequently allowing the pooling of the libraries, after which fragments are paired-end sequenced using NGS Illumina sequencing, e.g. on the Illumina NEXTSEQ 1000 (Illumina, Inc,, San Diego, CA). This approach resulted in sequenced fragments that all share the sequence of the gene specific primer (e.g., RpoB6-SPA-seq-F 1652 primer) followed by sequences that vary in length and nucleotide composition. Sequences derived from the same microorganisms will be identical except for the length of the sequenced fragment, which will vary in function of the distance between the gene specific primer (e.g., RpoB6-SPA-seq-F1652 primer) annealing site and the end of the mcfDNA fragment. A similar protocol was followed for creating SPA fragments from the 16S rRN A gene using the 16S-seq-V4-R primer.

[00152] The analysis of the SPA fragment sequences included the following steps:

1. Adaptors and primers are trimmed from the sequences.

2. Using DADA2, an open-source software used for fast and accurate sample inference from amplicon data with single-nucleotide resolution (Callahan et al, 2016), the following steps are performed: a. Heads are filtered based on read quality. b. The remaining reads of different lengths are deduplicated. c. Reads are error-corrected using a parametric error model. d. Error-corrected reads are resolved to Amplicon Sequence Variants ( ASVs).

3. ASVs of the RpoB6-F1652 primer or the 16S-V4-R primer are aligned to either the rpoB or 16s gene database using the basic local alignment search tool (BLAST, Altschul et al, 1990).

[00153] The database of bacterial rpoB genes was initially created by downloading their nucleotide sequences from the PATRIC database (Wattam et al, 2014) using the version available January 2021. If more than one (incomplete) rpoB gene was found for the same genome, we accepted the longest one, and rejected the shorter one(s). We confirmed for several instances our assumption that multiple rpoB genes in a single strain represented assembly errors, since each bacterium contains only one rpoB gene per genome. Genes were rejected if the genome had no taxonomy or if the gene was not annotated as “DNA-directed RNA polymerase beta subunit (PC 2.7.7.6)”. We evaluated all annotation rejections and found none that seemed to be rejected incorrectly. After January 2021. any new genome added to our genome database is searched for a rpoB gene by annotation, “DNA-directed RNA polymerase beta subunit (EC 2, 7.7.6)” and if found, its nucleotide sequence is added to the database of bacterial rpoB genes. These genomes come from PATRIC and NCR I (National Center for Biotechnology Information; https://www.iwbi.nhn.nih.gov/). Our curated database of bacterial rpoB genes contains 59,069 unique nucleotide sequences as of November 2021. For 16S sequences the .16S_ribosomal_RNA database was downloaded from NCBL

[00154] The lengths of the ASV fragments for the RpoB6-F1652 primer and the 16S- V4-R primer are shown in Figure 4 and Figure 5, respectively. The SPA fragment length distributions are in line with the size distributions of mcfDNA. These fragments are slightly shorter than the lengths reported by Burnham et al (2016) as the primer annealing site was trimmed from the sequences.

[00155] Table 2 is a sample of alignment results for the RpoB6-FI652 primer-based SPA fragment sequences, while Table 3 provides a sample of alignment, results for the 16S- V4-R primer-based SPA fragment sequences. The presented alignments were required to have an identity of at least 90% across 90% of the bases of the query. E-values represent the probability of the alignment occurring by chance. In the sample results for the 16S-V4-R primer, a SPA fragment as short as 40 nucleotides was aligned with confidence of an E-value of L94E-I4 against the I6S rRNA gene of strain 034. Both 16S rRNA gene and rpoB gene derived SPA fragments were found for Flavobacterium, Staphylococcus, and Ps eudom mas . ruble 2: Sample alignment results for RpoB6-F 1652 SPA fragments to the rpoB gene database. For each fragment, the percentage of identity, fragment length and alignment length to a reference genome are indicated. E-values represent the probability of the alignment occurring by chance.

Table 3: Sample alignment results of 16S-V4-R SPA fragments to the 16S rRNA gene database. For each fragment, the percentage of identity, fragment length and alignment length to a reference genome arc indicated. E- values represent the probability of the alignment occurring by chance

EXAMPLE 2

Primer selection and SPA protocol based on the rpoB gene as phylogenetie marker.

[00156] As a representative example, the SPA sequencing approach was successfully demonstrated for design of a rpaB gene specific SPA primer. A total of 50,569 unique rpoB gene sequences were downloaded from the PATRIC database (Wattam et al, 2014) using the version available in January 2021. RpoB gene sequences were identified based on their annotation as “DNA-directed RNA polymerase beta subunit (EC 2.7.7.6)”.

[00157] A subset of 50 rpoB gene sequences, representative for a broad range of phylogenetically distinct eubacterial reference microbes, were initially aligned by clustalW to identify conserved n ucleotide regions of the rpoB gene, resulting in the identi ficat ion of several conserved regions as primer candidates. This included the rpoB gene regions 1327- 1352, 1528- 1550, 1690-1709, 3766-3788 and 3808-3830, as well as the two regions identified by Ogier et al (2019), region 1630-1652 and region 2039-2063. The positions of the regions are based on the nucleotide sequence of the Escherichia coli rpoB gene.

[00158] Using the SPA algorithm, the 50,569 unique rpoB genes sequences were aligned to these conserved regions to determine their nucleotide compositions. The conserved nucleotide sequences of the rpoB gene regions 1327-1352, 1528-1550 and 1690-1709 are provided in Figures 7 A, 8 and 9 as representative examples. In Table 4, the average sequence variances for the primer candidates is shown, with al I primer candidates having a similar score, making them all primer leads. Subsequently, the estimate of adjacent region conservation was calculated as described above. For each region, which represents a putative primer annealing site, the variance is shown for 25, 50, 75, 100 or 200 nucleotides (nt) upstream (5’) or downstream (3’) of the beginning or end of the sequence of the conserved region. The results are summarized in Table 4 and show that the nucleotide sequence upstream of the conserved region 1327- 1352 is the most variable, as indicated by the lowest average variance scores of 0.0667 for both the 25 nucleotide-long and 50 nucleotide-long regions. This variability is also shown in Figures 10A and 10B, where the variance score for the 75 nucleotides upstream or downstream of the conserved region 1327-1352 has been plotted. Figures IGA and 10B also show the conservation of the nucleotides in the region 1327-1352, as well as the positions of the proposed degenerate primers RpoB i-Rl 327 and RpoB 1-F 1352, respectively. The sequences of the degenerate primers RpoB1-R1327 and RpoB1-F1352 are shown in Table I, The identification of a hypervariable DNA region in the rpoB gene upstream of the conserved region 1327-1352 was unexpected, as it falls outside of the region that has previously been identified and used for RpoB gene amplicon sequencing (Ogier ei al, 2019).

[00159] To select primers with the least risk for nonspecific annealing to human genomic DNA, the number of putative annealing sites of the proposed degenerate primer sequences to the human genome sequence (Reference: GCF 000001405.40 GRCh38.p 14 genomic.fna) with increased number of allowed mismatches is determined. Results for the degenerate primers 16S-V3-F, 16S-V4-R, 16S-V6- R, RpoB1-R1652, RpoB7-R2039 and RpoB-R1327 arc shown in Table 5. A primer should not have zero or one mismatch, and ideally no more than 10 instances of two mismatches with the human genome. Based on the results from this analysis, the primer 16S-V3-F showed an unexpectedly high number of putative annealing sites to the human genome, especially compared to the 16S-V4-R primer that also targets the V3-V4 region of the 16S rRNA gene and is, based on this result, considered unsuitable for SPA fragment sequencing.

Table 4: Average sequence variance for the primer regions and the regions upstream or downstream of candidate primer annealing regions recognizing conserved rpoB gene sequences. For each region adjacent to the primer region, the variance is shown for 25, 50. 75, 100 or 200 nucleotides (nt) upstream (5’) or downstream (3’) of the beginning or end of the primer annealing sequence. The variance score is cal culated as the average of the variance of the percentage of the nucleotides adenine, guanidine, cytosine and thymine at each position of the rpoB gene. A lower number is indicative for more variance, while a higher number is indicative for less variance and a more conserved DNA sequence. The maximum theoretical variance score for a region is 0.25 (would represent a 100% conserved DNA region). Regions with a variance score <0.1 are highlighted. The coordinates of the regions recognized by the primers are based on the nucleotide sequence of the Escherichia cali rpoB gene.

Table 5: Number of hits for primers to the human genome. For each primer, the number of hits with zero, one or two mismatches are presented. The number of hits was determined based on homology to the nucleotide sequence both DNA strands (+ and --- strand) of the human chromosome (Reference: GCF 000001405.40 GRCh38.p14 genomic, fna). [00160] We subsequently analyzed the minimal length of the variable regions required to have sufficient sequence-based phylogenetic resolution for species level identification, while keeping in mind the size of mcfDNA fragments of approximately 40-100 bp as determined by Burnham et al (2016) and Rassoulian Barrett et al (2020). To do so we calculated the numbers of unique SPA fragments with length of 25, 50, 75, 100 and 200 nucleotides for the regions located downstream of the annealing sites for the RpoB 1-R 1327 and RpoB7-R2039 primers, and upstream of the RpoBl- F1352 and RpoB6~F1652 primers, respectively. The results are presented in Figure 1 1 and show that the region upstream of the annealing site for primer RpoB1-R1327 consistently provided a higher number of unique SPA fragments compared to the other three primers, especially in the size range up to 75 nucleotides. For 50 nucleotide length, 20,919 unique SPA fragments could be generated for the upstream region. Based on the results presented in Table 4, Figures 10 A and 10B, and Figure 11 the degenerate RpoB- R l 327 primer, which recognizes the conserved rpoB gene region 1327-1352 and al lows for the generation of SPA fragments from the region upstream of the primer annealing site, was selected to validate in silico the Single Point Amplicon (SPA) fragment sequencing protocol for the rpoB gene and was added to our SPA primer repository.

[00161 ] The RpoB 1-R. I 327 primer, which recognizes the rpoB gene sequence between positions 1327 - 1352 (positions based oaths Escherichia call rpoB gene sequence) and targets the region upstream of the primer annealing site, was validated in silico for the phylogenetic resolution of 50 nucleotide Single Point Amplification (SPA) fragments as described in EXAMPLES 3 to 9. In EXAMPLES 7 and 9 we also validated the R.poB6-R 1630 primer, which recognizes the rpoB gene sequence between positions 1630 - 1652.

EXAMPLES 3 to 9

[00162] To analyze their phylogenetic resolution, sequences of 50 nucleotide long SPA fragments located upstream of the RpoB1-R1327 priming site were generated in silico and analyzed on the genus or species level. Based on the size range of mcfDNA in blood of approximately 40-100 bp (Burnham et al, 2016) it very likely that SPA fragments of approximately 50 nucleotides can be obtained, this in addition to a small number of larger fragments. In EXAMPLES 3 to 9 we demonstrate that 50 nucleotide long SPA fragments provide sufficient phylogenetic resolution to distinguish a wide range of clinically relevant pathogenic bacteria at the species level. To further increase the resolution, we also validated in EXAMPLES 7 and 9 the RpoB6-R1630 primer, which recognizes the rpoB gene sequence between positions 1630 — 1652.

[00163] It should be noted that since we compare the SPA fragments for strain identification against a deduplicated database, the number of strain s found for a SPA fragment represents the number of distinct rpoB gene sequences that share a common SPA fragment.

EXAMPLE 3

SPA fragment sequences for identification of Mycobacterium species.

[00164] Tuberculosis (TB) is an infectious disease for which cfDNA sequencing based diagnostics seems very promising. Clinical recognition of TB is hampered by its long latency and nonspecific presenting symptoms. In addition, people who have received the Bacillus Calmette- Guerin (BCG) vaccine cannot be tested for active TB using routine skin test screening (https:/ Avwwxdc.gov/tb.dopic/testingTestingbcgvaccinated.htn). Of the estimated 10.4 million active TB cases occurring worldwide in 2016, it is estimated that 40% remained either undiagnosed or unreported, in large part due to inadequate diagnostics. Etiological diagnosis is typically delayed when reliant solely on the acid-fast bacillus (ABB) culture method, while invasive biopsies are often necessary to cultivate the pathogen from deep-seated infections. For an early diagnosis of tuberculosis, researchers have established several targeted Mycobacterium tuberculosis mcfDNA assays (PCR-based methods) to determine the presence of infection by detecting Mycobacterium tuberculosis mcfDNA in blood and urine specimens (Fernandez-Carballo et al, 2019). More recently, the performance of deep plasma mcfDNA sequencing was evaluated in patients with tuberculosis infection, including the direct detection in a series of cases of Invasive Mycobacterium chimaera infection (Nomura et al, 2019), providing accurate noninvasive microbiologic confirmation in approximately 4 days, which was more than one month faster than standard AFB culture method. Similarly, other successful applications in diseases such as opportunistic Mycobacterium avium or Mycobacterium tuberculosis infections in HIV/AIDS patients (Zhou et al, 2019) and aneurysms infected by Mycobacterium bovis due to Bacille Calmette-Guerin (BCG) instillation (Vudatha et al, 2019) demonstrate that mcfDNA analysis provides a promising, less-invasive diagnostic and monitoring tool for TB. Unfortunately, due to the need for costly deep NGS sequencing, mcfDNA sequencing is not feasible for routine and large-scale screening for TB. Therefore, there is an unmet need for high-resolution, high-throughput and low-cost detection of Mycobacterium tuberculosis and other disease-causing Mycobacterium strains, something SPA fragment sequencing can deliver. As such, TB and the detection of Mycobacterium species represents an important application for SPA fragment sequencing-based detection.

[00.165] To evaluate its application for the reliable detection of TB and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting Mycobacterium species, 50 nucleotide long SPA fragments located upstream of the RpoBl-Rl 327 priming site were generated in silieo for Mycobacterium strains. The results are presented in Table 6.

Table 6; Overview of the sequences of 50 nucleotide SPA fragments generated in silico for Mycobacterium species. For each SPA fragment, the Mycobacterium species and the number of strains is indicated. The SPA fragments representing 456 Mycobacterium strains are reported. A^wtec/ermm-specific (My) SPA fragments received a unique numerical identifier for reference in further analysis. Unique SPA fragments with a single Mycobacterium species hit were not reported.

[00166] The 50 nucleotide SPA fragments were found to be highly distinctive for clinically relevant Mycobacterium species, including Mycobacterium tuberculosis, Mycobacterium avium, Mycobacterium chimaera and Mycobacterium leprae. For instance, the dataset included 290 Mycobacterium tuberculosis plus Mycobacterium tuberculosis subsp. ajricanum strains that could be identified by two distinct SPA fragments, SPA fragments Myl and My2. SPA fragment Myl identified 291 strains. In addition to 286 Mycobacterium tuberculosis strains and one Mycobacterium tuberculosis subsp. africanum strain, this fragment was also present in three Mycobacterium canettii strains and one Mycobacterium orygis strain, both members of the Mycobacterium tuberculosis complex and very closely related to Mycobacterium tuberculosis.

[00167] The similarities between the strains identified by SPA fragments Myl and My 2 was analyzed using whole genome-based Average Nucleotide Identity (Arahal, 2014). The results are presented in Figure 12 and show that representative strains of Mycobacterium tuberculosis, Mycobacterium tuberculosis subsp. qfricanum and Mycobacterium orygis shared ANI values of 100%, indicating that they represent identical species. The ANI values of these strains with the three Mycobacterium caneltii strains ranged between 98% to 99%, similar to the ANI values shared between the three Mycobacterium caneltii strains, indicating that all strains are very closely related and that Mycobacterium caneltii is likely a Mycobacterium tuberculosis subspecies, as confirmed by the shared SPA fragment Myl.

[00168] Mycobacterium avium strains, which can cause serious infection in immune compromised patients, such as HIV AIDS patients, are identified by two distinct SPA fragments, My8 and My9. In addition to recognizing four Mycobacterium avium strains, SPA fragment My9 also identified two metagenome assembled genomes (MAG), Mycobacterium MAC_011194_8550 and Mycobacterium MAC_ 080597_8934. Based on the specificity of this fragment for Mycobacterium avium it is assumed that the two MAGs are representatives of Mycobacterium avium, as was confirmed by whole genome-based ANI analysis (Figure 13).

[00169] The 97 strains belonging to Mycobacterium abscessus and Mycobacterium abscessus subsp. Massiliense could be identified by five distinct 50 nucleotide SPA fragments (My 3 to My7), with no other species being identified. Unique SPA fragments also identified the clinically relevant species Mycobacterium chimaera (My 10) and Mycobacterium leprae (Myl 1).

[00170] A few SPA fragments identified multiple distinct Mycobacterium species. For instance, eight strains of 'Mycobacterium conceptionense. Mycobacterium formitum (2 strains), Mycobacterium neworleansense, Mycobacterium nonchromogenicum, Mycobacterium vitifteris, Mycolicibaclerium boenickei, and Mycobacterium senegalense shared the common 50 nucleotide SPA fragment My 17. Except for Mycobacterium nonchromoge/ticum, these strains all belong to the Myeolicibacteriuin gen. nov. clade) and are very closely related (Gupto et al, 2018). It is generally accepted in the field that AN ! values around 95% correspond to the 70% DNA-DNA hybridization cut-off value, which is widely used to delineate archaeal and bacterial species (Arahal, 2014), Whole genome-based ANI analysis (Figure 14) showed that these strains indeed represent distinct species. Similar, closely related members of the emended genus Mycobacterium (“tuberculoste-rimiae” clade) represented by Mycobacterium liflandii, Mycobacterium marinum, Mycobacterium pseudoshottifii, Mycobacterium shottsii and Mycobacterium ulcerous shared the common 50 nucleotide SPA fragment My 18. In this specific case, the ANI values between the various strains ranged between 97% to 100%, confirming that they are closely related and part of the same genus Mycobacterium (“tuberculorix-rimiae”) clade. This group (My 18) is also highly distinct from the Mycobacterium strains identified by the SPA fragment My 17, with ANI scores of 74% to 75% (Figure 14). Increasing the length of the SPA fragments to 75 nucleotides did not significantly improve their phylogenetic resolution.

[00171 ] These results show that, unexpectedly, despite their relatively short size, sequences of 50 nucleotide long SPA fragments covering the region upstream of the RpoBI- R1327 primer annealing site allow for high resolution phylogenetic identification of Mycobacterium at the species or clade level (as summarized in Table 7), including the clinically relevant species. This shows the importance and potential of SPA fragment sequencing as a new approach for high-throughput TB screening, based on the (early) detection and identification of infectious Mycobacterium species using mcfDNA from peripheral blood and/or urine samples.

Tabic 7: Summary of rhe Mycobacterium (My) specific SPA fragments as phylogenetic identifiers at the species or clade level. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl-R 1327 primer annealing site.

EXAMPLE 4

SPA fragment sequences for the detection of bacterial pathogens associated with pulmonary infection risks i n cystic fibrosis patients.

[00172] Cystic fibrosis (CF), the most common autosomal genetic disease in North America affecting I ;2000 Caucasian individuals, is characterized by chronic lung malfunction, pancreatic insufficiencies and high levels of chloride in sweat. Its high mortality index is evident when lung and spleen are affected, but other organs can also be affected. The persons affected die by progressive bronchiectasis and chronic respiratory insufficiency, CF patients will see a succession of lung inflammation by opportunistic pathogenic bacteria. During the first decade of li fe of CF patients, Staphylococcus aureus and Hemophilus influenzae are the most common bacteria, but in the second and third decade of life, Pseudomonas aeruginosa is the prevalent bacterium. Other important infectious bacterial pathogens associated with pulmonary infection risks in cystic fibrosis patients include Nontuberculous Mycobacteria (NTM) and Burltholderia cepacia (for review, see Coutinho ei al, 2008), Therefore, there is an unmet need for high-resolution, high-throughput and low-cost detection of opportunistic pathogenic bacteria in CF patients, something SPA fragment sequencing can provide. The same is generally true for patients having a compromised immune system.

[00173] Mycobacterium species: The most common NTM infecting CF patients are Mycobacterium abscessus (identified by SPA fragments My 3 to My7), Mycobacterium avium (identified by SPA fragments My8 and My9), and Mycobacterium (paraflntracellulare (identified by SPA fragments Myl3), with Mycobacterium abscessus the NTM more likely associated with the disease, all of which can be identified by their unique SPA fragments (see Table 7).

[00174] Staphylococcus aureus: This is usually the first pathogen to infect and colonize the airways of CF patients. This microorganism is prevalent in children and may cause epithelial damage, opening the way to the adherence of other pathogens such as Pseudomonas aeruginosa. To evaluate its application for the reliable detection of chronic infection in CF patients by Slaphylocoecus aureus and related species, and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting Staphylococcus species, 50 nucleotide long SPA fragments located upstream of the RpoBl- R1327 priming site were generated in silica for Staphylococcus strains. The results are presented in Table 8.

Table 8: Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Staphylococcus aureus species. For each SPA fragment, the Staphylococcus species and the number of strains is indicated. The SPA fragments representing 545 Staphylococcus aureus and strains that shared their SPA fragment are reported. Staphylococcus <ww,s-specific (Sa) SPA fragments received a unique numerical identifier for reference in further analysis. Unique SPA fragments with a single Staphylococcus aureus species hit were not reported.

[00175] Based on the SPA fragment sequences, four mixed clusters were identified, each with their unique 50 nucleotide fragment (Tab le 8). that contained Staphylococcus aureus. Whole genome-based ANI analysis on representative members of these four clusters revealed that they grouped in three highly distinct species (Figure 15).

[00176] ANI group I, comprised of strains identified by SPA fragments Sal and Sad. With the exception of a single Staphylococcus hyicus strain, the 521 strains identified by Sa l and Sa2 were all Staphylococcus aureus. Since the Staphylococcus hyicus strain had a 98% ANI score with the Staphylococcus aureus strains, similar to the score between Staphylococcus aureus strains, it also belongs to this species (Arahal, 2014). This confirms that SPA fragments Sal and Sa2 are specific for the identification of Staphylococcus aureus strains.

[00177] ANI group II, comprised of strains identified by SPA fragment Sa3. These strains had been previously identified as Staphylococcus argenteus and Staphylococcus aureus. Since these strains had ANI scores of 87% to 88% with the ANI group I Staphylococcus aureus strains, they represent a different species (Arahal, 2014), most likely Staphylococcus argenteus. Thus, SPA fragment Sa3 seems to be specific for the identification of Staphylococcus argenteus strains.

[00178] ANI group III, comprised of strains identified by SPA fragment Sa4. These strains had been previously identified as Staphylococcus schweltzeri and Staphylococcus aureus. Since these strains had ANI scores of 88% to 89% with the ANI group I Staphylococcus aureus strains and 92% with the ANI group II Staphylococcus argenteus strains, they represent a different species (Arahal, 2014), most likely Staphylococcus schweitzeri. Thus, SPA fragment Sa4 seems to be specific for the identification of Staphylococcus schweitzeri strains.

[00179] Despite their relatively short size, 50 nucleotide long SPA sequencing fragments covering the region upstream of the RpoBl-R1327 primer annealing site allow for high resolution phylogenetic identification of Staphylococcus at the species level (as summarized in 'fable 9), including the clinically relevant species Staphylococcus aureus and Staphylococcus argenteus.

Table 9: Summary of the Staphylococcus aureus (Sa) specific SPA fragments as phylogenetic identifiers at the species level The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB1-R1327 primer annealing site.

[00180] Pseudomnas aeruginosa; This species is part of the normal microbial population of the respiratory tract, where it is an opportunistic pathogen in CF patients. Pseudomonas aeruginosa causes infections in more than 50% of CF patients, especially in adult CF patients, as infection has been shown in 20% CF patients 0-2 years old while in 81% in adult groups (>18 years old). To evaluate its application for the reliable detection of chronic infection in CF patients by Pseudomonas aeruginosa and related species, and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting species, 50 nucleotide long SPA fragments located upstream of the RpoBl-R1327 priming site were generated in silica for Pseudomonas aeruginosa strains. The results are presented in Table 10.

Table 10: Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Pseiukmwnas aeruginosa species. For each SPA fragment, the Pseudomonas species and the number of strains is indicated . The SPA fragments representing 564 Pseudomonas aeruginosa and strains that shared their SPA fragment are reported. Pseudomonas ^erwgiwosa-speciftc (Pa) SPA fragments received a unique numerical identifier for reference in further analysis. Unique SPA fragments with a single Pseudomonas aeruginosa species hit were not reported.

[00181 ] Based on the SPA fragment sequences, four clusters were identified, each with their unique 50 nucleotide fragment (Table 10), that contained Pseudomonas aeruginosa. ANI analysis on representative members of these four clusters revealed that they grouped in three highly distinct species (Figure 16). Based on the results presented in Figure 16, two major ANI groups can be distinguished for the Pseudomonas strains identified by the SPA fragments Pal , Pa2 and Pa4.

[00182] ANI group I, which is comprised of strains identified by SPA fragments Pal and Pa2, represents Pseudomonas aeruginosa. Based on their ANI scores of 98% to 99%, the Pseudomonas fluorescens strain NCTCT0783 and the Acinetobacter baumannii strain 4300STDY7045820 were previously misclassified and represent Pseudomomas aeruginosa strains. The only strain identified by SPA fragment Pa2 that fell outside of ANI group I was Pseudomonas psychrotolerans strain DSM 15758. This should cause no problem as this species, which grows at lower temperature than P. aeruginosa, is not clinically relevant.

[00183] ANI group III, which is comprised of strains identified by SPA fragments Pa4. This group, which includes three Pseudomonas strains, is based on its ANI score (76% to 78%) distinct from the Pseudomonas aeruginosa strains identified by SPA fragments Pal and Pa2.

[00184] Thus, despite their relatively short size, sequences of 50 nucleotide long SPA fragments covering the region upstream of the RpoB1-R1327 primer annealing si te allow for high resolution phylogenetic identification of Pseudomonas aeruginosa at the species level (as summarized in Table 11).

Table 11 : Summary' of the Pseudomonas aeruginosa (Pa) specific SPA fragments as phylogenetic identifiers at the species level. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl -R 1327 primer annealing site.

[00185] Burkholderia cepacia complex (SCO: A bacterial complex with twenty genomic species (genomovars): genomovar 1 (B, cepacia), II (B. mullivorans), III (B. cenocepacia), EV (B. stabilis), V (B. vietnamiensis), VI (B. dolosa), VII (B. ambifaria), VIII (B. anthina), IX (B. pyrrocinia), and more recently B. stagna/is, B. territorii, B. ubonensis, B. eontaminans, B, seminalis, B. metallica, B. arboris, B. lata, B. latens, B. pseudomultivorans, and B. diffusa was reported by Depoorter et al (2016). Infected CF patients show high levels of BCC in the salivary fluid, with transmission rates, prognosis and mortality being distinctly characteristic for each genomovar, as are the treatment strategies. Of the over 20 formally named species within the complex, Burkholderia multivorans (genomovar II) and Burkholderia cenocepacia (genomovar III) together account for approximately 85-97% of all BCC infections in CF (Savi et al, 2019). To evaluate its application for the reliable detection of chronic infection in CF patients by Bulkholderia cepacia and related BCC complex species, and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting Burkholderia species, 50 nucleotide long SPA fragments located upstream of the RpoBl-R1327 priming site were generated in silico for Burkholderia strains. The results are presented in Table 12.

Table 12: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for members of the Burkholderia cepacia complex. For each SPA fragment, the Burkholderia species and the number of strains is indicated. The SPA fragments representing 567 Burkholderia cepacia complex members (marked in bold) and related strains that shared their SPA fragment are reported. Burkholderia cepacia complex-specific (Bcc) SPA fragments received a unique numerical identifier for reference in further analysis. Unique SPA fragments with a single Burkholderia cepacia complex species hit were not reported. * Indicates species whose name and has not been officially accepted.

[00186] Based on the SPA fragment sequences, seven clusters were identified, each with their unique 50 nucleotide fragment (’fable 12), that contained Burkholderia cepacia. ANI analysis on representative members of various clusters defined by the SPA fragments Bccl (, Bed and Bcc2, Bccl and Bcc3, and Bed. Bcc6 and Bcc7, revealed that 50 nucleotide SPA fragments fail to phylogenetically distinguish between tire Burkholderia cepacia complex strains, hi addition, a very limited number of strains that fall outside the Burkholderia cepacia complex were found to have similar SPA fragments. ANI analysis confirmed that these strains, such as Parabacleroides strains found to have SPA fragment Bed , were not misclassifie

[00187] To address the lack of phylogenetic resolution of 50 nucleotide SPA fragments for Burkholderia cepacia complex strains, larger SPA fragments were analyzed. Increasing the SPA fragment length to 75 nucleotides had only a minor effect on the phylogenetic resolution. For instance, the extended 75 nucleotide version of SPA fragment Bed identified 479 strains, with the major difference being the removal of five Paraburkholderia strains. However, increasing the SPA fragment length to 100 nucleotides resulted in the breakup of the SPA fragment Bed group with increased phylogenetic resolution that allowed for differentiation between several species belonging to the Burkholderia cepacia cluster. Since we expect to get for each species a limited number of SPA fragments with sizes around 100 nucleotides, as we showed in EXAMPLE 1, SPA fragment sequencing should allow for classification of Burkholderia cepacia cluster species with sufficient phylogenetic resolution. This is shown in Table 1.3 and Table 14 for the strains initially identified by the 50 nucleotide SPA fragment Bed .

Table 13: Overview of the sequences of 100 nucleotide SPA fragments generated in silica for members of the Burkholderia cepacia complex that share the SPA fragment Bed . For each SPA fragment, the Burkholderia species and the number of strains is indicated. The SPA fragments representing 471 Burkholderia cepacia complex members (marked in bold) and related strains that shared their SPA fragment are reported. Burkholderia cepacia complexspecific (Bee) SPA fragments received a unique numerical identifier for reference in further analysis. * Indicates 100 nucleotide SPA fragments. Unique SPA fragments with a single Burkholderia cepacia complex species hit were not reported. ($) indicates that Burkholderia thailandensis was incorrectly identified as this species, and as shown in Figure 17 represents a new Burkholderia species.

[00188] Using 100 nucleotide long SPA sequencing fragments covering the region upstream of the RpoBl-R.1327 primer annealing site significantly increased the resolution for phylogenetic identification of Burkholderia cepacia complex species, as is summarized in Table 14.

Table 14: Summary of the Burkholderia cepaeia complex (Bee) specific SPA fragments and their phylogenetic resolution for strains that that share the SPA fragment Bed. The SPA fragments are 100 nucleotides in length and cover the region upstream of the RpoB1 -R1327 primer annealing site. ($) indicates the presence of species from outside the Burkholderia cepacia complex.

[00189] Burkholderia pseudomallei group: Most members of the Burkholderia pseudomallei group including Burkholderia mallei, Burkholderia oklahomensis and Burkholderia pseudomallei are considered pathogenic. Table 15 shows that two unique SPA fragments, Bpnil and Bpm2, reliably identified these clinically relevant species. Burkholderia thailandensis, also a member of the Burkholderia pseudomallei complex, is generally considered nonpathogenic. Burkholderia thailandensis could be identified by its own unique SPA fragment, Bpm3. This result, which was also confirmed by the ANI analysis of Figure 16, further demonstrates the clinical relevance of the SPA as an important method for (early) detection and identification of Burkholderia species at the level of their major pathogenic complexes using mcfDNA from peripheral blood samples. The results form Figure 17 also show that the Burkholderia thailandensis strain, previously shown to have SPA fragment Bed , was incorrectly identified as this species, but instead represents a new Burkholderia species.

Table 15: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for members of the Burkholderia pseudomallei group. For each SPA fragment, the Burkholderia pseudomallei group species and the number of strains is indicated. The SPA fragments representing 137 Burkholderia pseudomallei group members ( marked in bold) and related strains that shared their SPA fragment are reported, Burkholderia pseudomallei group-specific (Bpm) SPA fragments received a unique numerical identifier for reference in further analysis. Unique SPA fragments with a single Burkholderia pseudomallei group species hit were not reported. [00190] Haemophilus irifluefizae: This species usually infects younger CF patients. For example, in Brazil, 20.4% of CF children between 6 and 12 years old are infected by Haemophilus influenzae. This bacterium hyper-mutates, which can be related to its resistance to antibiotics, making treatment more difficult. To evaluate its application for the reliable detection of chronic infection in CF patients by Haemophilus influenzae and related species, and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting Haemophilus influenzae species, 50 nucleotide long SPA fragments located upstream of the RpoB 1 -R 1327 priming site were generated in silieo for Haemophilus influenzae strains. The results are presented in Table 16.

Table 16: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for Haemophilus influenzae species. For each SPA fragment, the Haemophilus influenzae species and the number of strains is indicated. The SPA fragments representing 136 Haemophilus influenzae strains and Haemophilus strains that shared their SPA fragment are reported. Haemophilus influenzae -specific (Hi) SPA fragments received a unique numerical identifier for reference in further analysis. Unique SPA fragments with a single Haemophilus influenzae species hit were not reported. [00191] The species identified by the SPA fragments Hi1, H2. Hi6 and H17 were further analyzed by AN I, which resulted in the identification of two distinct ANI groups (Figure 18);

ANI group I. comprised of strains identified by SPA fragments Hi2 and Hi 6, represents the Haemophilus parainfluenzae strains. If also shows thatPartewreZ/aceae HGM20799, which has an ANI score of 94% to 95% with the other strains in this cluster, should be reclassifies as Haemophilus parainfluenzae.

ANI group II, comprised of strains identified by SPA fragments Hi 1 and Hi7, represents the Haemophilus influenzae strains. It also shows that the Haemophilus aegyplius strain, which has ANI scores of 97% with the other strains in this cluster, should be reclassifies as Haemophilus influenzae, The

Haemophilus haenwlylicus strain, which was identified by SPA fragment Hi7, seems to be an outlier in this group with an ANI score of 89% with the other strains in this cluster.

[00192] Compared to other species, the ANI scores between members of the same ANI group are relatively low, around 95% instead of 98% to 99%. This might reflect the hyper-mutation phenotype of members of the genus Haemophilus. Overall, sequences of 50 nucleotide long SPA fragments covering the region upstream of the RpoBl-R1327 primer annealing site allow for high resolution phylogenetic identification at' Haemophilus influenzae and Haemophilus parainfluenzae at the species level (as summarized in Table

Table 17: Summary of the Haemophilus (para)infhtenzae (Hi) specific SPA fragments as phylogenetic identifiers at the species level. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB 1 -R1327 primer annealing site.

[00193] Overall, SPA fragments are capable of high resolution phylogenetic identification of opportunistic pathogenic bacteria frequently found to cause infections in OF patients. As such, SPA fragment sequencing represents a powerful tool to evaluate infections in CF patients as their treatment, including the selection of antibiotics, depends on the correct identification of the infectious species.

EXAMPLE 5

SPA fragment sequences to identify o rnistic bacterial pathogens linked to sepsis.

[00194] Opportunistic pathogens of clinical relevance, including Pseudomonas aeruginosa, Mycobacterium abcessus, and Staphylococcus aureus, have been found as the cause of sepsis in patients with compromised immune systems. The successful use of SPA fragments for the high-resolution phylogenetic identification of these species has been described in EXAMPLES 3 and 4.

[00195] Streptococcus species, including 5. pneumonia, S. pyogenes and 5. intermedins are also frequently found as opportunistic pathogens in patients with compromised immune systems, such as HIV.'AIDS patients, organ transplant patients or cancer patients undergoing chemotherapy. In addition, other clinically relevant Streptococcus species such as Streptococcus gallolyticus, Streptococcus macedonicus, Streptococcus pasteurianus and Streptococcus equinus, have been linked to cancer. Therefore, there is an unmet need for high- resolution, high-throughput and low-cost detection of opportunistic pathogenic Streptococcus species, something SPA fragment sequencing can provide. To evaluate its application for the reliable detection in peripheral blood of opportunistic pathogenic bacteria leading to sepsis by Streptococcus species, and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting Streptococcus species, 50 nucleotide long SPA fragments located upstream of the RpoBl-R1327 priming site were generated in silica for Streptococcus strains. The results are presented in Table 18.

Table 18: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for Streptococcus species. For each SPA fragment, the Streptococcus species and the number of strains is indicated. The SPA fragments representing 1 ,712 Streptococcus species and strains that shared their SPA fragment are reported. Smyjtococi'u^-specifie (St) SPA fragments received a unique numerical identifier for reference in further analysis. Unique SPA fragments with at least seven Streptococcus strain hit were reported, with the exception of Streptococcus intermedins and Streptococcus gallofyticus sub.sp. gallofyticus

[00196] Overall, 50 nucleotide long SPA sequencing fragments covering the region upstream of the RpoBl-R1327 primer annealing site allow for high resolution phylogenetic identification of Streptococcus at the cluster or species level (as summarized in Table 20). For instance, unique SPA fragments were able to phylogenetically identify Streptococcus mutans, the cause of dental cavities; Streptococcus suis, a pathogen in pigs that can cause severe systemic infection in humans; Streptococcus agalactiae and Streptococcus equi, the causative agent of strangles which is the most frequently diagnosed infectious disease of horses; and Streptococcus parauberis, an important fish pathogen. When multiple species were identified by the same 50 nucleotide SPA fragment, whole genome-based ANI analysis on representative members was used to confirm the results based on the SPA fragments. Representative examples are shown in Figures 19 to 23, where ANI analysis was used to confirm the phylogenetic specificity of the Streptococcus SPA fragments.

[00197] The ANI results shown in Figure 19 confinn that Streptococcus dysgal'actiae is only identified by SPA fragment St2, while Streptococcus pyogenes is identified by SPA fragments St2, St3 and St4. Both species are veiy closely related and belong to the Lancefield Group A Streptococci.

[00198] Similarly, based on ANI results, the SPA fragments Stl , Sts. St9, St 10, Stl 1 and Stl 2 can be used to identify bacterial strains belonging to the Streptococcus initis. Streptococcus pneumoniae and Streptococcus pseudopneumoniae cluster. Members of this duster have previously been referred to as the viridans group streptococci (VGS), Streptococcus mills group, and based on their ANI analysis, group together. A second group of strains, identified by the SPA fragments St19, St20 and St22, represents bacterial strains previously identified as Streptococcus mitis and Streptococcus oralis ( Figure 20). Based on their ANI score, these strains belong to a different group than those identified by the SPA fragments St 1 , St8, St9, St10, St 11 and Stl 2. As most of the strains identified by SP/\ fragments St19, St20 and St22 were identified as Streptococcus oralis, with ANI scores between the Streptococcus mitis and Streptococcus oralis strains of this AN.I group being similar (91% to 94%) and significantly different from the ANI scores of the Streptococcus mitis/Streptococcus pneumoniae/Strepwcoccus pseudopneumoniae group members (86%), it is concluded that these strains are Streptococcus oralis. The results shown in Figure 20 also confirm that the strains identified by SPA fragment St21 are Streptococcus gordonii and Streptococcus oligofermenlans. Based on their ANI scores of 95% to 96% these two oral Streptococcus species are very closely related.

[00199] As shown in Figure 21, Streptococcus anginosus, Streptococcus conslellalus and Streptococcus intermedins form a cluster of tightly related strains. Based on ANI analysis, three ANI groups can be distinguished: ANI group I, comprised of Streptococcus anginosus strains identified by SPA fragments Stl 4 and St 17; ANI group III, comprised of Streptococcus intermedius strains identified by SPA fragments St 14, Stl5 and St 16; and ANI group II, comprised of Streptococcus anginosus, Streptococcus constellatus and Streptococcus intermedius strains all identified by SPA fragment St 14. Based on their whole genome-based ANI scores, the ANI group II strains belong to the same species and are distinct from the Streptococcus anginosus, and Streptococcus intermedius strains of ANI groups I and II, and most likely represent Streptococcus constellatus.

[00200] The whole genome-based ANI analysis for the Streptococcus equinus, Streptococcus salivarius. Streptococcus thermophilus and Streptococcus vestibularis strains identified by SPA fragments St30, St31 and St32 is shown in Figure 22 and identifies three distinct ANI groups: ANI group I and II representing Streptococcus thermophilus strains and Streptococcus vestibularis strains, respectively, identified by SPA fragment St30; and ANI group 111 representing Streptococcus salivarius strains identified by SPA fragments St30, St.31 and St32, Based on the ANI score it can also be concluded that Streptococcus equinus strain FDAARGOS_251, identified by SPA fragment St30, was misidentified and represents a Streptococcus salivarius strain. [00201] Streptococcus gallolyticus subsp. gallolyticus (formerly known as Streptococcus bovis type I) has recently been recognized as the main causative agent of septicemia and infective endocarditis in elderly and immunocompromised persons. It also has been strongly associated to colorectal cancer (CRC; defined as carcinomas and premalignant adenomas) (Boleij et al, 201 1; Pasquereau-Kotula et al, 2018). Several previous studies failed to clearly attribute an association between Streptococcus bovis and CRC; this can. however, be explained by the lack of a proper distinction between Streptococcus bovis type I (Streptococcus gallolyticus strains), type II. I (Streptococcus infantarws) and type II.2 (Streptococcus gallolyticus subsp. macedonicus and Streptococcus gallolyticus subsp. pasteurianus), with Streptococcus bovis type I being prevalently associated to CRC, and to a lesser extend Streptococcus bovis type II.2 (Abdul amir et al, 201 1 ), The phylogenetic resolution of 50 nucleotide SPA fragments allowed to discriminate between Streptococcus infantarius (SPA fragment St28) anti Streptococcus gallolyticus (SPA fragments St33 and St35) strains. Therefore, SPA fragment sequencing represents a promising approach for CRC screening based on the presence of Streptococcus galloly ticus strains (Streptococcus bovis type I and 11.2 ) in peripheral blood. The whole genome-based AN ! analysis presented in Figure 23 show that the three subspecies Streptococcus gallolyticus subsp. gallolyticus, Streptococcus gallolyticus subsp. Macedonicus and Streptococcus gallolyticus subsp. pasleuriamts are very closely related. It also shows that Streptococcus gallolyticus subsp. gallolyticus NCTC8133 should be classified as Streptococcus equinus.

[00202] Since Enterococcus faecalis and Enterococcus /aecium also belong to the Lancefield group D “Streptococci” (Table 20), the phylogenetic resolution of their 50 nucleotide SPA fragments was determined (Table 19). SPA fragments Efl and Ef2 were found to be specific for Enterococcus faecalis , while SPA fragments Ef3 and Ef4 were found to be specific for Enterococcus Jaeciutn. The results of the whole genome-based AN1 analysis, shown in Figure 24, confirmed the separate clustering of these two species. It also confirmed the misclassification of the Streptococcus pneumoniae and the Enterococcus lactis strains listed in Table 18 among the Enterococcus faecalis and Enterococcus faecium strains identified by SPA fragments Ef2 and Ef3. Based on their ANI scores, these strains should be reclassified as Enterococcus faecalis and Enterococcus faecium strains, respectively.

Table 19: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for Enterococcus faecaiis and Enterococcus faecium strains. For each SPA fragment, the Enterococcus faecaiis and Enterococcus faecium species and the number ofstrahis is indicated. The SPA fragments representing 266 Enterococcus species and strains that shared their SPA fragment are reported. Enterococcus faecaiis and Enterococcus faecium-specific (Ef) SPA fragments received a unique numerical identifier for reference in further analysis. Unique SPA fragments with a single Enterococcus faecaiis or Enterococcus faecium species hit were not reported.

Table 20: Summary of the phylogenetic specificity of 50 nucleotide SPA fragments generated upstream of the RpoBl-R1327 primer annealing site for clinically relevant Streptococcus species (SPA fragments Stl to St35) and Enterococcus species (SPA fragments Efl to Ef4). Where applicable, the Lancefield group (Lancefield, 1933) or the viridans group streptococci (VGS) subgroup are indicated, as well as the standard of care antibiotic treatment for infections caused by specific Streptococcus species.

[00203] These results show that, unexpectedly, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoBl-R.1327 primer annealing site allow for high resolution phylogenetic identification of clinically relevant Streptococcus and Enterococcus at the species or group level (as summarized in Table 20), thus providing an important method for ( early) detection and identification of these infectious species using mcfDNA from peripheral blood samples. The high phylogenetic resolution of the SPA fragments can be directly linked to the standard of care for the most appropriate antibiotic regime to treat infections by Streptococcus and species, further demonstrating the clinical relevance of SPA fragment sequencing.

[00204] In addition, clinically relevant Streptococcus species including Streptococcus gallolyticus, Streptococcus macedonieus, Streptococcus pasteurianus and Streptococcus equinus have been linked to cancer. Therefore, the detection of Streptococcus species in peripheral blood is important for detection and prognostics of various types of cancer, as will also be discussed in EXAMPLE 7.

[00205] Furthermore, the analysis of EXAMPLE 5 shows the promise of SPA fragment sequencing as a new approach for assessing the risk of sepsis in immune compromised individuals, based on the (early) detection and identification of infectious and opportunistic pathogenic bacterial species using mcfDNA from peripheral blood samples.

EXAMPLE 6

SPA fragment sequences to identify opportunistic bacterial pathogens originating from the oral cavity.

[00206] The oral cavity represents a source of opportunistic pathogenic bacteria that can have significant health implications when entering the body, Porphyromonas gingivalis is an example of an oral pathogen that has received a lot of attention. Not only is this bacterium the cause of gingivitis (Socransky et al. 1998; Chen et al, 2018), but several studies have implicated this bacterium in Alzheimer's disease (Dominy et al, 2019; Kanagasingam et al, 2020). Therefore, in the fight against Alzheimer's disease there is an unmet need for higli- resolution, high-throughput and low-cost early detection of this bacterium in peripheral blood, something SPA fragment sequencing can provide. To evaluate its application for high- resolution detection of Porphyromonas gingivalis in peripheral blood, saliva or stool to complement risk screening for developing Alzheimer's disease, and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting Porphyromonas gingivalis strains, 50 nucleotide long SPA fragments located upstream of the RpoB1-R1327 priming site were generated in silica for Porphyromonas strains. The results are presented in Table 21 ,

Table 21 : Overview of the sequences of 50 nucleotide SPA fragments generated in silica tor Porphyromonas gingivalis strains and related species. For each SPA fragment, the Porphyromonas species and the number of strains is indicated. The SPA fragments representing 63 Porphyromonas species and related strains are reported. Porphyromonas (gingivalis) -specific (Pg) SPA fragments received a unique numerical identifier (for reference in further analysis.

[00207] As shown in Table 21, the 50 nucleotide SPA fragments generated in silico for Porphyromonas gingivalis strains and related species distinguish Porphyromonas at the species level, as was also confirmed by whole genome-based ANI analysis (Figure 25). The AN I analysis shows that the Porphyromonadaceae identified by the SPA fragments Pg3, Pg4 and Pg9 form a new ANI group. ANI analysis also confirms that the Porphyromonas endodontalis and Propionibacterium acidifaciens strains, identified by SPA fragment Pg7, are very closely related (100% ANI score) and therefore represent the same species. These results show that unexpectedly, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoBl-R1327 primer annealing site allow for high resolution phylogenetic identification of clinically relevant Porphyromonas at the species level (Table 21 ), including Porphyromonas gingivalis, thus providing an important method for its (early) detection using mcfDNA from peripheral blood, saliva and stool samples. This shows the importance of SP A fragment sequencing as a new approach as part of risk screening for Alzheimer's disease based on the (early) detection and identification of Porphyromonas gingivalis species .

[00208] Prevolella are bacteria that inhabit many parts of the body. Although common in the gut microbiome, if found elsewhere, they may be a sign of infection. Prevotella oris represents an example of an opportunistic pathogenic bacterium that has been associated with several serious oral and systemic infections, Prevotella oris can been identified in clinical specimens by bacterial culture and biochemical tests, which are generally unreliable (Riggo and Lennon, 2007). Therefore, there is an unmet need for high-resolution, high-throughput and low-cost early detection of this bacterium in peripheral blood, something SPA fragment sequencing can provide. To evaluate its application for the reliable detection in peripheral blood of Prevotella species, and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting Prevotella strains, 50 nucleotide long SPA fragments located upstream of the RpoBl-R.1327 priming site were generated in silico for Prevotella strains. The results are presented in Table 22.

Table 22: Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Prevotella species. For each SPA fragment, the Prevotella species and the number of strains is indicated. The SPA fragments representing 63 Prevotella species strains are reported, Prevotella-specific (Pr) SPA fragments received a unique numerical identifier for reference in further analysis.

[00209] As shown in Table 22, the 50 nucleotide SPA fragments generated in silico for Prevotella strains distinguish Prevotella at the species level. These results show that unexpectedly, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoBl-R1327 primer annetiling site allow for high resolution phylogenetic identification of clinically relevant PrevoieUa at the species level (Table 22), including PrevoieUa oris, thus providing an important method for its (early) detection as an infecting pathogen using mcfDNA from peripheral blood samples.

EXAMPLE 7

SPA f bacteria.[inked to junior rnicrobioings md th for cancer detection and progression monitoring.

[00210] Several clinically relevant bacteria have been identified as playing a key role in the onset and progression of cancer, such as S’riryjroeocczcv bovis type I (Streptococcus gallolyticus strains) which has been associated with CRC (see Example 5). Therefore, the use of SPA fragment sequencing for screening of peripheral blood or stool of cancer patients for the presence of bacteria as biomarkers for the detection, monitoring of disease progression, prognostics for survival and minimal residual disease, will provide important information complementaiy to customary blood biopsy- and stool-based detection and monitoring approaches that use cfDNA and focus on the methylation and mutation footprints in specific genetic loci as tumor biomarkers.

[0021 1 ] Contrary to PCR-based detection methods that monitor for the presence of specific bacteria, SPA fragment sequencing provides an “open” diagnostics approach to detect any bacterium based on the presence of its mcfDNA in peripheral blood. Due to its high phylogenetic resolution, SPA fragment sequencing can be used to identify novel microbiome signatures in blood and stool as biomarkers for the (early) detection of cancer. Ones these signatures have been identified and validated as cancer-relevant biomarkers, SPA fragment sequencing is ideally positioned as a novel high-resolution, high-throughput and low-cost approach for population screening, e.g. adults between the ages 45 to 85, with a focus on (early) detection. In what follows, examples are provided for SPA fragments as biomarkers to detect and monitor the progression of cancer based on the presence of microbial signatures characterized by bacteria that have been associated with specific cancers and their developmental stage.

[00212] Risk screening for esophageal cancer: Esophageal cancer is the eighth most common cause of cancer deaths worldwide. Tannere.Ua forsythia and Porphyrontonas gingivalis, both of which have been implicated in periodontal diseases as part of red complex of periodontal pathogens, have been found to be associated with an increased risk of esophageal cancer (Malinowski et al, 2019). As shown in Table 21 of EXAMPLE 6, Porphyromonas gingivalis strains can be specifically identified by SPA fragment Pgl . To evaluate its application for the reliable detection in peripheral blood, saliva and stool of Tamierel/ci forsythia to complement risk screening for developing esophageal cancer, and to analyze the discriminatory power of SPA fragment sequencing for high resolution phylogenetic identification of infecting Tan tiefella forsythia strains, 50 nucleotide long SPA fragments located upstream of the RpoBl~R1327 priming site were generated in silica for Tannerella forsythia strains. The results are presented in Table 23.

Table 23: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for Tannerella forsythia and the related species Tannerella oral. For each SPA fragment, the Tannerella species and the number of strains is indicated. The SPA fragments representing 10 Tannerella strains are reported. Tannerella (forsythia)-specific ( Tf) SPA fragments received a unique numerical identifier for reference in further analysis.

[00213] As shown in 'fable 23, the 50 nucleotide SPA fragments generated in silico for Tannerella strains distinguish between Tannerella forsythia and the related species Tannerella oral. These results show that unexpectedly, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering tire region upstream of the RpoBI -R1327 primer annealing site allow for high resolution phylogenetic identification of the clinically relevant species Tannerella forsythia (Table 23). Therefore, SPA fragments for Tannerella forsythia and Porphyromonas gingivalis can be used as biomarkers using mcfDNA from peripheral blood, saliva and stool samples for the risk profiling and (early) detection of esophageal cancer.

[00214] Risk screening for precanceraus colonic polyps: The common commensal bacterium, nontoxigenic Bacteroidesfragilis (NTBF), is enriched in patients with precancerous colonic polyps. NTBF isolated from polyps is enriched in genes involved in EPS biosynthesis, which may allow for its increased ability to activate the immune system and cause inflammation (Kordahi et al, 2021). Therefore, there is an unmet need for high-resolution, high- throughput and low-cost early detection of this bacterium in peripheral blood and stool samples, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the early detection of Bacteroides fragilis as an indicator species lor the presence of precancerous colonic polyps, 50 nucleotide long fragments located upstream of the RpoB I-R1327 priming site were generated in silieo for Bacteroides fragilis strains. The results are presented in Table 24.

Table 24: Overview of the sequences of 50 nucleotide SPA fragments generated in silieo for Bacteroides fragilis and related species. For each SPA fragment, the Bacteroides species and the number of strains is indicated. The SPA fragments representing 80 Bacteroides fragilis strains and related species are reported. Bacteroides /ragz/w-specific (Bf) SPA fragments received a unique numerical identifier reference in further analysis. [00215] As shown in Table 24, the 50 nucleotide SPA fragments generated in silica for Bacteroides fragilis strains and related species distinguish Bacteroides fragilis at the species level, as was also confirmed by whole genome-based ANI analysis presented in Figure 26. Whole genome-based ANI analysis shows that the Bacteroides fragilis strains identified by the SPA fragments Bf2 and BD form an ANI group distinct from the Bacteroides fragilis identified by the SPA fragment Bfl and might represent a different species or subspecies. AN I analysis also confirms that the Bacteroides cellulyticus strain, identified by SPA fragment BD, is nearly identical (100% ANI score) to Bacteroides fragilis strains and therefore represent the same species. Overall, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoB1-R1327 primer annealing site allow for high resolution phylogenetic identification of clinically relevant Bacteroides fragilis at the species and likely subspecies level (Table 24; Figure 26), thus providing an important method for its (early) detection using mcfDNA from peripheral blood samples. This shows the importance of SPA fragment sequencing as a new approach for the detection of precancerous colonic polyps based on the (early) detection and identification of Bacteroides fragilis species.

[00216] Risk screening fbr precancerous siomack ulcers: Stomach ulcers, caused by Helicobacter pylori, are a cause for stomach cancer when left untreated. Therefore, there is an unmet need for high-resolution, high-throughput and low-cost early defection of this bacterium in peripheral blood and stool, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the early detection of Helicobacter pylori as an indicator species for the presence of stomach ulcers and potentially early-stage stomach cancer, 50 nucleotide long fragments located upstream of the RpoBl-R1327 priming site were generated in silico for Helicobacter pylori strains. The results are presented in Table 25.

Table 25; Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Helicobacter pylori. For each SPA fragment the number of Helicobacter pylori strains is indicated. The SPA fragments representing 6 Helicobacter pylori strains are reported. Helicobacter pylori-specific (Hp) SPA fragments received a unique numerical identifier for reference in further analysis.

[00217] As shown in Figure 27, whole genome-based AN I analysis reveals the presence of at least five select subspecies of Helicobacter pylori , with the strains identified by SPA fragment Hpl breaking up in three AM I groups. Overall, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoB 1 -R1327 primer annealing site allow for high resolution phylogenetic identification of the clinically relevant species Helicobacter pylori (Table 25). Therefore, SPA fragments for Helicobacter pylori can be used as biomarkers using mcfDN A from peripheral blood and stool samples for the risk profiling and (early) detection of precancerous stomach ulcers. The blood antibody test, a blood test to evaluate whether your body has made antibodies to Helicobacter pylori bacteria, is commonly used to determine if a patient is either currently infected or has been infected in the past with this bacterium. The advantage of SPA fragment sequencing is that it will only detect an active infection by Helicobacter pylori.

[00218] Women 's health risk screening: Chlamydia trachomatis, a bacterium which is commonly transmitted sexually, is the major cause of mucopurulent cervicitis, pelvic inflammatory disease, tubal factor infertility, and ectopic pregnancy. Thus, the healthcare costs due to complications caused by Chlamydia trachomatis are enormous. [00219] Cervical cancer is the most common cancer in women worldwide. Infection with Chlamydia trachomatis greatly increases the risk of cervical cancer ( Anttila et al, 2001 ). Although infections with oncogenic strains of human papillomavirus remain the prime cause of cervical cancer, coinfections with some strains of Chlamydia trachomatis and Neisseria gonorrhoeae seem to contribute to that risk and the severity of the disease, especially highgrade squamous intraepithelial cervical lesions (De Abreu et al, 2016). This finding is important because chlamydia, though frequently asymptomatic, is one of the most common sexually transmitted diseases and can be treated with appropriate antibiotics, hi the United States, between tour million and eight million new cases of chlamydia are reported yearly.

[00220] Neisseria gonorrhoeae is a bacterial pathogen responsible for gonorrhea and various sequelae that tend to occur when asymptomatic infection ascends within the genital tract or disseminates to distal tissues. Like Chlamydia trachomatis. Neisseria gonorrhoeas is an important sexually transmitted pathogen and a major cofactor in HIV- 1 infection. Global rates of gonorrhea continue to rise, facilitated by the emergence of broad-spectrum antibiotic resistance that has recently afforded the bacteria 'superbug' status. Therefore, there is an unmet need for high-resolution, high-throughput and low-cost early detection of these bacteria in peripheral blood and vaginal smear samples, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the early detection of Chlamydia trachomatis and Neisseria gonorrhoeae as indicator species for women's health issues including the risk to develop cervical cancer, 50 nucleotide long fragments located upstream of the RpoB1-R1327 priming site were generated in silico for Chlamydia trachomatis and Neisseria gonorrhoeae strains. The results are presented in Table 26 and Table 27. Tabic 26: Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Helicobacter pylori. For each SPA fragment the number of Chlamydia trachomatis strains is indicated. The SPA fragments representing 27 Chlamydia trachomatis strains are reported. Chlamydia trachomatis-specific (Ct) SPA fragments received a unique numerical identifier for reference in further analysis.

[00221 ] These results indicate that unexpectedly, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoB1-R1327 primer annealing site allow for high resolution phylogenetic identification of the clinically relevant species Chlamydia trachomatis (Table 26).

Tabic 27: Overview of the sequences of 50 nucleotide SPA fragments generated in siiico for Neisseria species. For each SPA fragment, the Neisseria species and the number of strains is indicated. The SPA fragments representing 167 Neisseria strains and related species are reported. TVeissma-specific (Ne) SPA fragments received a unique numerical identifier for reference in further analysis.

[00222] Except for SPA fragments Nel and Ne4, the A'mjrerm-specifie (Ne) SPA fragments were found to be species specific. The major combined group, identified by SPA fragment Nel, was formed by Neisseria gonorrhoeas and Neisseria meningitidis strains. Neisseria meningitidis (meningococcus) causes significant morbidity and mortality in children and young adults worldwide through epidemic or sporadic meningitis and or septicemia.

[00223] To improve the phylogenetic resolution of SPA fragment sequencing for Neisseria species, 50 nucleotide long fragments located upstream of the RpoB6-R 1630 priming site were generated in silieo for Neisseria strains. As shown in Table 4, the region upstream of the RpoB6-R1630 priming site has less sequence variance than the region upstream of the RpoBl-R1327 priming site. However, we found that this region provided a high degree of phylogenetic resolution of several cl 12Iinically important bacteria, including strains belonging to the genus Neisseria, An overview of the phylogenetic resolution of RpoB6-R1630-based SPA fragment sequencing for Neisseria species is provided in Table 28.

Table 28: Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Neisseria species from the region upstream of the RpoB6-R1630 priming site. For each SPA fragment, the Neisseria species and the number of strains is indicated. The SPA fragments representing 169 Neisseria strains and related species are reported. AWs.vm«-specific (Ne) SPA fragments received a unique numerical identifier or reference in further analysis,

[00224] As shown in Table 28, SPA fragments generated in silico for Neisseria species from the region upstream of the RpoB6-Rl 630 priming site allowed to distinguish with high phylogenetic resolution between Neisseria gonorrhoea# and Neisseria meningitidis strains. The practical implications of using an alternative primer annealing site or a combination of two primers that target different phylogenetic identifier regions are discussed in EXAMPLE 9.

[00225] Overall, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoB6~R1630 primer annealing site allow for high resolution phylogenetic identification of the clinically relevant species Neisseria gonorrhoea# (Table 28). Therefore, SPA fragments for Chlamydia trachomatis and Neisseria gonorrhoeae can be used as biomarkers using mcfDNA from peripheral blood and/or vaginal smear samples for the risk profiling and (early) detection of women's health issues related to these bacteria including the risk to develop cervical cancer.

[00226] Prognostic correlations with the microbiome of breast cancer subtypes: There are four subtypes of breast cancer (BC) that are based on the status of the estrogen receptor, progesterone receptor, and human epidermal growth (Her2) expression in cancerous breast cells. As shown by Banerjee et al (2021), the subtypes of BC have specific viromes and microbiomes, with estrogen receptor positive (ER+) and triple negative (TN) tumors showing the most and least diverse microbiomes, respectively. These specific microbial signatures allowed successful discrimination between the different BC subtypes. Furthermore, Banerjee et al (2021 ) demonstrated correlations between the presence and absence of specific microbes in BC subtypes with the clinical outcomes. Therefore, there is an unmet need for high- resolution, high-throughput and low-cost early detection of bacteria associated with breast cancer subtypes in peripheral blood, something SPA fragment sequencing can provide. [00227] TN BC ( 15-20% of BC patients) is the most aggressive of all the BCs, is non- responsive to treatment, is highly angiogenic, highly proliferative, and has the lowest survival rate, TN breast cancer showed decreased microbial diversity and increased levels of Aggregatibacter species; significant levels of this species were not detected in other BC types. Therefore, there is an unmet need for high-resolution, high-throughput and low-cost early detection of this bacterium in peripheral blood, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the detection of Aggregatibacter species as indicator and prognostics species for TN breast cancer, 50 nucleotide long fragments located upstream of the RpoB1-R.1327 priming site were generated in silica for Aggregatibacter strains. The results are presented in Table 29. Tabic 29: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for Aggregatibacter species. For each SPA fragment, the Aggregatibacter species and the number of strains is indicated. The SPA fragments representing 37 Aggregatibacter strains and related species are reported. Aggregatibacter- specific (Ag) SPA fragments received a unique numerical identifier for reference in further analysis.

[00228] The results presented in Table 29 showed that the Aggregatibacter species could be identified by their unique SPA fragments. This was further confirmed by whole genome AN’I analysis (Figure 28).

[00229] The whole genome-based ANI results in Figure 28 confirmed that Aggregatibacter actinomycetemcomitans could be identified by SPA fragments Agl and Ag3; that Aggregatibacter aphrophilus could be identified by SP A fragments Ag2 and Ag4; and that Aggregatibacter segnis could be identified by SPA fragment Ag5 (see also Table 29). Overall, these results show that unexpectedly, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoBl-R.1327 primer annealing site allow for high resolution phylogenetic identification of clinically relevant Aggregatibacter species. Therefore, SPA fragments for Aggregatibacter can be used as biomarkers using nicfDNA from peripheral blood and/or saliva samples for the risk profiling and (early) detection of TN breast cancer, as well as other cancers. For instance, a prospective populationbased nested case-control study demonstrated that the presence of Porphyromorias gingival is or Aggregatibacter actinomycetemcomitans in the oral cavity was indicative of increasing the risk of pancreatic cancer (Chandra and McAllister, 2021 ).

[00230] Prognostic correlations with the microbiome of pancreatic cancer: Pancreatic, cancer, particularly pancreatic ductal adenocarcinoma (PDAC), is an aggressive disease with a poor prognosis. Chandra and McAllister (2021 ) pointed out the importance of microbial biomarkers lor risk prognosis lor pancreatic cancer. Risk factors for pancreatic cancer included periodontal disease and oral microbial dysbiosis, with abundances of Porphy romonas gingivalis, Aggregatibacter actinomycetemcomitans, Neisseria elongate and Streptococcus mills as indicator species. As discussed previously, 50 nucleotide SPA fragments covering the region upstream of the RpoB 1 -R 1327 primer annealing site can be used to successfully identify these species.

[00231] Of specific interest is the tumor microbiome composition of PDAC patients, as it bolds clues for their treatment options and long-term survival. Geller et al (2017) reported the presence of bacteria in human PDACs and demonstrated that intra-tumoral Gamma-proteobacteria, among the most common bacteria detected in human pancreatic tumors, reduce the efficacy of chemotherapeutic drugs like gemcitabine, which these bacteria can metabolize into its inactive form via their cytidine deaminase. Thus, one application of SPA fragment sequencing would be to link phylogenetic identification to metabolic strain models, thereby predicting the impact of the tumor microbiome on drug metabolism and efficacy.

[00232] Riquelme et al (2019) profiled intra-tumoral bacteria from patients with resected PDAC and compared short-term and long-term survivors. Long-term survivors had greater intra-tumoral microbial «-di versity than did those who died of the disease within 5 years after resection. Overall tumor microbial characterization revealed a microbial composition similar to the one in human PDAC previously described by Geller et al (2017), but unique enrichment in the following microbes was found in tumors from long-term survivors: Pseudoxanthamonas, Streptomyces, Saccharopolyspora and Bacillus clausii, the last two species have documented immunomodulatory functions that might play a role in slowing down the disease progression. Therefore, there is an unmet need lor high-resolution, high-throughput and low-cost early detection of these bacteria in peripheral blood, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the detection of Pseudaxanthomonas, Streptomyces, Saccharopolyspora and Bacillus clausii species as indicator and prognostics species as prognostics for long terra survival of PDAC patients, 50 nucleotide long fragments located upstream of the RpoBl-R1327 priming site were generated in silico for Pseudaxanthomonas, Streptomyces, Saccharopolyspora and Bacillus clausii strains. Unique SPA fragments were found able to identify Pseudoxanthomonas and Streptomyces at the genes level, and Saccharopolyspora and Bacillus clausii at the species level. The results for Bacillus clausii are presented in Table 30.

Table 30: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for Bacillus clausii strains. For each SPA fragment, the Bacillus clausii species and the number of strains is indicated. The SPA fragments representing 14 Bacillus clausii strains and related species are reported. Bacillus cfausii-specific (Bel) SPA fragments received a unique numerical identifier for reference in further analysis.

[00233] These results show that overall, unexpectedly, despite their relatively short size, 50 nucleotide long SP A fragment sequences covering the region upstream of the RpoBl- R1327 primer annealing site allow for high resolution phylogenetic identification of clinically relevant Pseudoxanthomonas, Sireplomyces, Saccharopolyspara and Bacillus clausii strains. Therefore, SPA fragments can be used as biomarkers using mcfDNA from peripheral blood samples for the risk profiling and prognostics for long-term survival of PDAC patients.

[00234] Prosjnostic correlations with the microbiome of lung cancel': Lung cancer is the most common cancer, excluding nonmelanoma skin cancer, and the most common cause of cancer-related death in the world, with approximately 1.8 million diagnoses and 1.6 million dea ths per year. Peters et al (2019) pointed out the importance of microbial biomarkers for risk prognosis for lung cancer, observing that greater abundance of family Koribacieraceae in normal long tissue was associated with increased recurrence- free survival (RFS) and long-term disease-free survival (DFS), whereas greater abundance of family Lachnospiraceae, and genera Faeealibacterium and Rumiuococcus (from Ruminococcaceae family), and Roseburia and Riuninocaccus (from Lachnospiraceae family) were associated with reduced RFS and DFS. Taxa associated only with RFS (P < 0.05) included family S24-7 (increased RFS), and family Bacleroidaceae and genus Bacteroides (reduced RFS). Taxa associated only with DFS (P < 0.05) included family Sphingomonadaceae and genus Sphingomonas (increased DFS), and family Ruminococcaceae (reduced DFS). However, this study was performed using 16S rRNA gene sequencing and lacked the phylogenetic resolution to identify biomarker species at the species level. The 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoBl-R1327 primer annealing site allow for the high resolution phylogenetic identification at the species level of the clinically relevant bacteria associated with the prognosis for recurrence-free survival (RFS) and long-term disease- free survival (DFS) of lung cancer patients. SPA sequencing is therefore well positioned to monitor disease progression and prognosis for lung cancer patients.

[00235] Risk screenings for gastrointestinal tumors: Fusobacterium spp. is important in the development and progression of gastrointestinal tumors. In line with this, Poore el al (2020) showed that the Fusobacterium genus was overabundant in primary tumors compared to normal solid-tissue. Furthermore, pan-cancer analyses also showed an overabundance of Firsobacierium when comparing all broadly-defined gastrointestinal (Gl) cancers against non-Gl cancers in both primary tumor tissue and adjacent normal solid-tissue, pointing to Fusobacterium species as a biomarker for Gl cancer. Therefore, there is an unmet need for high-resolution, high-throughput and low-cost early detection of this bacterium in peripheral blood and stool samples, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the early detection of Fusobacterium species as biomarker for the risk to develop gastrointestinal cancer, 50 nucleotide long fragments located upstream of the RpoB1-R1327 priming site were generated in silico for Fusobacterium species. The results are presented in Table 31.

Table 31: Overview of the sequences of 50 nucleotide SPA fragments generated zn silica for Fusobacterium species. For each SPA fragment, the Fusobacterium species and the number of strains is indicated. The SPA fragments representing 73 Fusobacterium strains and related species are reported. Fusobacterium-specific (Fs) SPA fragments received a unique numerical identifier for reference in further analysis.

[00236] As shown in Table 31 , the 50 nucleotide SPA fragments generated in silica for Fusobacterium strains mostly allowed to distinguish Fusobacterium at the (sub)species level, as was also confirmed by whole genome-based ANI analysis. The following exceptions were observed: In addition to identifying Fusobacteriutn nucleatum subsp. polymorphum, SPA fragment Fs 1 also identified the closely related Fusobacterium canifelinum. Whole genomebased ANI analysis confirmed the similarity between these two species. In addition to identifying Fusobacterium hwasookii, SPA fragment Fs7 also identified the closely related Fusobacterium nucleatum subsp . polymorphism. Whole genome-based ANI analysis confirmed the similarity between these two species; it also showed that Fusobacterium nucleatum ChDC F 128 strain should be reclassified as Fusobacterium hwasookii. Whole genome-based ANI analysis also showed that Fusobacterium eipiinum and Fusobacterium gonidia/ormans, both identified by SPA fragment Fs15. represent the same species. A summary of the Fusobacterium species (Fs) specific SPA fragments as phylogenetic identifiers at the (sub)species level is provided in Table 32.

[00237] These results show that unexpectedly, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoBl-R1327 primer annealing site allow for high resolution phylogenetic identification of clinically relevant Fusobacterium at the (sub)species level (Table 32), thus providing an important method for its (early) detection using mcfDNA from peripheral blood and stool samples. This shows the importance of SPA fragment sequencing as a new approach as part of risk screening for broadly-defined gastrointestinal (Gl) cancers based on the (early) detection and identification of Fusobacterium species.

Table 32: Summary of the Fusobacterium species (Fs) specific SPA fragments as phylogenetic identifiers at the species level. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoBl-R1327 primer annealing site, * Whole genomebased ANI analysis indicates that these species are nearly identical.

[00238] Several studies successfully demonstrated that including the microbial footprint increases the specificity and sensitivity of screening tests for the detection of early- stage adenomas and carcinomas in colorectal cancer. For example, a metagenomics-based classification model, using abundance changes of Fusobacterium nucleatum ssp, vincentii and Fusobacterium nucleatum ssp, animalis, Peptostreptococcus stomatis and Pseudcmoeardia asaccharofytica in CRC patients versus healthy controls combined with standard CRC diagnostics improved CRC-detection sensitivity for the guaiac-based fecal occult blood test (gFOBT) by >45% (Zeller et al, 2014). A microbiota-based random forest model using abundance changes of Fusobacterium, Peptostreptococcus, Porphyromonas, Prevotella, Parvimonas, .Bacteroides and Gemella species complemented the fecal immunochemical test (FIT) (Baxter et al, 2016). The microbiota-based random forest model detected 91.7 % of cancers and 45.5 % of adenomas while FIT alone detected 75.0 % and 15.7 %, respectively. Of the colonic lesions missed by FIT, the model detected 70.0 % of cancers and 37.7 % of adenomas.

[00239] The present inventors confirmed that Peptostreptococcus siomatis and Pseudonocardia asaccharofylica can be identified by their single unique SPA fragments; that Panwm species, including Parvimonas oral and Parvimonas micra could be identified by a single SPA fragment; and that Gemella species, including Gemella morbillorum, Gemella haemolysans, Gemella palalicanis and Gemella sanguinis each bad their unique SPA fragment. Therefore, combining tumor-specific biomarkers (including mutational footprint, methylation footprint, and blood detection in stool) with the quantitative detection of biomarker microorganisms using SPA fragment sequencing at the species and subspecies level will significantly increase the sensitivity and specificity of colorectal cancer screening. In addition, a further application of the SPA sequencing method is that once unique SPA fragments have been identified that correlate with the detection of specific diseases and monitoring of their progression, the unique SPA fragment sequences can be used to develop species-specific screening assays as part of PCR-based diagnostic platforms.

[00240] In certain instances, disease phenotypes caused by bacteria will depend on specific metabolic properties; as a result, accurate disease detection, monitoring and prognostics will require additional functional insights besides phylogenetic identification and community composition. For example, a random forest-based model using abundance changes of Fusobaeterium nueleatum, Peptostreptococcus stomatis, Pseudonocardia asaccharotytica, Prevotella species, Parvimonas species, Gemella morbillorum and other bacteria, combined with gFOBT, improved the sensitivity/specificity of CRC detection (Thomas et al, 2019). This study also found that the choline trimethylaminelyase gene, which encodes Trimethylamine (TMA) synthesis from dietary quaternary amines (mainly choline and carnitine), was overabundant in the microbiomes of CRC patients (P = 0.001 ), identifying a relationship between gut microbiome choline metabolism and CRC. Trimethylamine (TMA) has previously been associated with atherosclerosis and severe cardiovascular disease. Importantly, SPA fragment sequencing provides the flexibility to address both phylogenetic identification and community functionality. For example, this is performed by selecting a degenerate primer that recognizes a conserved DNA region of a specific function, the same protocol outlined in Figures 2 and 3A is broadly applicable for SPA amplification and sequencing of -functional genes. Furthermore, phylogenetic and functional information can be obtained simultaneously by including both a degenerate primer that targets the phylogenetic identifier gene and a degenerate primer that targets the functional gene in the same reaction for the SPA fragment amplification step (Figure 2, step 4). We refer to this approach as multiplex SPA for the simultaneous detection of multiple targets in a single PCR reaction. In the specific case of colorectal cancer, a primer targeting the choline trimethylaminelyase gene can be combined with the RpoBl-R1327 primer for improved detection, monitoring and progression of adenomas and carcinomas.

EXAMPLE 8

SPA fragment sequences for the detection of infections caused by emerging pathogenic bacteria of clinical concern.

[00241 ] Risk screening for developing Clostridium difficile injection: Clostridium difficile is the leading cause of health-care-associated infective diarrhea. Due to increased use of antibiotics that disrupt the healthy gut microbiome, creating a niche for Clostridium difficile to thrive, the incidence of Clostridium difficile infection (CDI) has been rising worldwide with subsequent increases in morbidity, mortality, and health care costs. Asymptomatic colonization with Clostridium difficile is common and a high prevalence has been found in specific cohorts, e.g., hospitalized patients, adults in nursing homes and in infants. Therefore, there is an unmet need for high-resolution, high-throughput and low-cost early detection of this bacterium in peripheral blood stool samples, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the early detection of Clostridium difficile as biomarker for the risk to develop CDI, 50 nucleotide long fragments located upstream of the RpoBl-R1327 priming site were generated in silico for Clostridium difficile strains. The results are presented in Table 33.

Table 33: Overview of the sequences of 50 nucleotide SPA fragments generated in silieo for Clostridium difficile strains. For each SPA fragment, the number of Clostridium difficile strains is indicated. The unique SPA fragment representing 60 Clostridium difficile strains is reported. The Clostridium difficile-specific (Cd) SPA fragment received a unique numerical identifier for reference in further analysis.

[00242] The results in Table 33 show that Clostridium difficile strains can be identified by the highly specific SPA fragment Cdi, thus providing an important method for its (early) detection using mcfDNA from peripheral blood samples. This shows the importance of SPA fragment sequencing as a novel approach as part of risk screening, e.g. after surgery or prolonged treatment with broad spectrum antibiotics, for developing CDI based on the (early) detection and identification of Clostridium difficile in peripheral blood and/or stool samples. Acinetobacter baumannii is an opportunistic bacterial pathogen primarily associated with hospital-acquired infections. "The recent increase in incidence, coupled with a dramatic increase in the incidence of multidrug-resistant (MDR) strains, has significantly raised the profile of this emerging opportunistic pathogen. Therefore, there is an unmet need for high-resolution, high-throughput and low-cost early detection of this bacterium in peripheral blood, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the early detection of Acinetobacter baumannii as biomarker for the risk to developing a hospital-acquired infection from this pathogen, 50 nucleotide long fragments located upstream of the RpoB1-R1327 priming site were generated in silieo for Acinetobacter baumannii strains. The results are presented in Table 34.

Table 34: Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Acinetobacter baumannii strains and related species. For each SPA fragment, the A cinetobacter species and the number of strains is indicated. The SPA fragments representing 506 Acinetobacter baumannii strains and related species are reported. Acinelobacler baumannii- specific (Ab) SPA fragments received a unique numerical identifier tor reference in further analysis.

[00244] As shown in 'fable 34, the 50 nucleotide SPA fragments generated in silico for Acinetobacter baumannii strains, especially SPA fragment Abl. largely allowed to distinguish Acinetobacter baumannii at the species level. However, several SPA fragments identified both Acinelobacier baumannii and related species, as well as some unexpected strains including Klebsiella pneumonia strains identified by SPA fragment Ab 1. To clarify this result, whole genome-based ANT analysis was performed on selected Acinelobacier baumannii strains and representatives from related species that were identified by the same SPA fragments. Where available, the genomes sequences of the Acinelobacier species type strains were included in this analysis, of which the results are shown in Figures 29, 30 and 31 . A total of eight ANI groups were identified:

ANI group I, which contains the strains identified by SPA fragment Abl (Figure 29). This group included representatives of the 346 Acinelobacier baumannii strains as well as three Klebsiella pneumoniae strains and an Acinetobacter calcoaceticus strain. Based on their ANI scores with Acinelobacier baumannii strains, including the type strain ATCC 17978, it was concluded that the Klebsiella pneumoniae strains and a Acinetobacter calcoaceticus strain had been misidentified and should be reclassified as Acinetobacter baumannii.

ANI group II, which contains Acinetobacter baumannii and Acinetobacter nosocomialis strains identified by SPA fragments Ab3 and Ab8 (Figure 29). Strains of ANI group II share very high ANI scores (>97%), indicating that they are the same species. Based on their low ANI scores with the ANI group I strains (91% to 92%), they represent a species closely related but distinct from Acinetobacter baumannii. Since the Acinetobacter nosocomialis type strain ANI was part of this group, the members of ANI group II should all be classified as Acinetobacter nosocomialis.

AN I group III, which contains Acinetobacter lactucae and Acinetobacter pittii strains identified by SPA fragment Ab2 (Figure 30). The group also contains an Acinetobacter pittii strain identified by SPA fragment Abl . Further analysis of the genome of this strain, which represents a metagenome assembled genome (MAG) of poor quality sequence, indicated that this MAG was highly contaminated and represented a chimeric assembly between Acinetobacter baumannii and Acinetobacter pittii. As such this MAG should be eliminated from the reference database. The group also contains Acinetobacter piitii strains identified by SPA fragment Ab6, as well as Acinetobacter baumannii strains identified by SPA fragments Ab1 and Ab6. Based on their whole genome-based ANI scores these strains are very similar to Acinelobacier pillii strains and should be reclassified as such.

ANI group IV, which contains closely related Acinetobacter calcoaceticus and Acinetobacter oleivorans strains identified by SPA fragments Ab2 and Ab4, as well as a strain identified by SPA fragment Ab4 that was misclassified as Acinetobacter baumannii (Figure 30),

ANI group V, which contains Acinetobacler baumannii and Acinetobacler radi ores is tens strains identified by SPA fragments Ab5 and Ab7 (Figure 31). Strains of ANI group V share very high ANI scores (>98%), indicating that they are the same species. Based on their low ANI scores with the ANI group I strains (75%), they represent a species different from Acinetobacter baumannii. Since the Acinetobacter radioresistens type strain DSM 6976 was part of this group, the members of ANI group V should ail be classified as Acinetobacter radioresistens.

ANI group VI, which contains Acinetobacter baumannii and Acinetobacter courvalinii strains identified by SPA fragment Ab 10 (Figure 31), Based on their low ANI scores with the ANI group I strains (77%), they represent a species distinct from Acinetobacter baumannii. and therefore, the Acinetobacter baumannii strains in this group should all be reclassified as Acinetobacter courvalinii. In addition, ANI group VI includes the Acinetobacter vivianii strains identified by SPA fragment Abb. ANI group VII, which contains Acinetobacter baumannii and Acinetobacter ursingii strains, including the Acinetobacter ursingii type strain DSM 16037, identified by SPA fragment Ab 13 (Figure 31). Based on their low ANI scores with the ANI group I strains (76%), they represent a species distinct from Acinetobacter baumannii, and therefore, the members of this group should all be reclassified as Acinetobacter ursingii.

ANI group VIII, which contains Acinetobacter baumannii and Acinetobacter variabilis strains identified by SPA fragment Ab9 (Figure 31). Based on their low ANI scores with the ANI group I strains (76%), they represent a species distinct from Acinetobacler baumannii, and therefore, the members of this group should all be reclassified as Acinetobacter variabilis.

[00245] Overall, these results confirm the phylogenetic resolution of 50 nucleotide SPA fragments to not only correctly identify Acinetobacter baumannii, but also point out strains that have been previously misclassified. A summary of the Acinetobacter baumannii strains and related species (Ab) specific SP.A fragments as phylogenetic identifiers at the species level is provided in Table 35. These results show that unexpectedly, despite their relatively short size, 50 nucleotide long SPA fragment sequences covering the region upstream of the RpoB 1 -R1327 primer annealing site allow for high resolution phylogenetic identification of clinically relevant Acinetobacter strains at the species level, thus providing an important method for their (early) detection using mcfDNA from peripheral blood samples. This shows the importance of SPA fragment sequencing as a new approach as part of risk screening for hospital acquired infections based on the (early) detection and identification of Acinetobacter species.

Table 35: Summary of the Acinetobacter baumannii strains and related species (Ab) specific SPA fragments as phylogenetic identifiers at the species level. The SPA fragments are 50 nucleotides in length and cover the region upstream of the RpoB 1 -Rl 327 primer annealing site.

EXAMPLE 9

Phylogenetic identification based on the combination of multiple SPA fragment sequences.

[00246] In addition to the previous examples, the example presented below demonstrates how the SPA fragment sequencing method is generalizable and adaptable to improve phylogenetic resolution in a targetable fashion, which is in.tbnned by the existing knowledgebase of sequence variation at the species and subspecies level. Just as a lens can be refocused, resolution can be redirected to identify new taxa and subspecies of interest.

[00247] To address a limited number of cases where 50 nucleotide long fragments located upstream of the RpoBl-R1327 priming site fail to identify bacteria at the genus or species level, the combination of two SPA fragments can be used to improve the phylogenetic resolution. In the example provided for the Enterobaeteriaceae, this is done by generating SPA fragments from two distinct regions of the rpoB gene and combining this information. However, the same can be achieved by combining the information of SPA fragments generated from two or more separate conserved housekeeping genes, including the prokaryotic genes coding for the DNA gyrase subunit B the chaperone protein (GroEL), the heat shock protein 60 (hsp60), the superoxide dismutase A protein ( wfr I ). the TU elongation factor (fw/), the 60 kDa chaperonin protein (cpn60), and DNA recombinase proteins (including recA, recE). Practically, the same protocol as outlined in Figure 2 would be used, except that two SPA primers would be included in the PCR reaction of Steps 4 and 5, resulting in the simul taneous generation of SPA fragments representing two regions for phylogenetic identification.

[00248] Screening for Enterobacteriaceae: The Enterobacteriaceae represents a group of often closely related bacteria, many of clinical importance. Key genera involve Escherichia, Shigella, Klebsiella, Salmonella and Serratia, many of which have been linked to sometimes life threatening and lethal infections, especially in immune compromised patients, including transplant patients where these bacteria are linked to post-transplant bloodstream infections. Graft versus Host Disease (GvHD), and increased mortality. Therefore, there is an unmet need for high-resolation, high-throughput and low-cost early detection of these bacteria in peripheral blood and other biopsy samples, something SPA fragment sequencing can provide. To analyze the discriminatory power of SPA fragment sequencing for the detection of Enterobacteriaceae. 50 nucleotide long fragments located upstream of the RpoBl-R.1327 priming site were initially generated in silico for members of the Enterobacteriaceae. The results for the two SPA fragments able to identi fy the largest number of strains are presented in Table 36.

Table 36; Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Enterobacteriaceae. For each SPA fragment, the Enterobacteriaceae species and the number of strains is indicated. The SPA fragments representing 1 ,989 Enterobacteriaceae strains. Enterobacteriacea-especific (Ent) SPA fragments received a unique numerical identifier for reference in further analysis.

[00249] As shown in Table 36, the 50 nucleotide SPA fragments generated in silico for strains belonging to the Enterobacteriaceae from the region upstream of the RpoB1-R13277 priming site failed to phyiogenetieally distinguish between strains on the genus level. This prompted us to eval uate if a combination of SPA fragments generated from two distinct regions of the rpoB gene would improve the phylogenetic identification of Enterobacteriaceae at the genus and species level. The results are presented in Table 37 and Table 38 for strains initially identified by SPA fragments Entl and Ent2, respectively.

Table 37: Overview of the sequences of 50 nucleotide SPA fragments generated in silico for

Entembacleriaceae. Strains were initially selected based on the presence of the 50 nucleotide SPA fragment End (see table 36), generated upstream of the RpoB1-R1327 priming site. Subsequently, 50 nucleotide SPA fragments were generated upstream of the RpoB6-R I 630 priming site. The sequences of these SPA fragments are presented and for each of these SPA fragments, the Enierobacteriiiceae species and the number of strains is indicated. SPA fragments identifying a single strain were left out. Enterobacteriaceae-specific (Ent) SPA fragments received a unique numerical identifier for reference in further analysis, with an asterisk symbol indicating that the SPA fragment was generated from the region upstream of the RpoB1-R-1630 priming site.

Table 38: Overview of the sequences of 50 nucleotide SPA fragments generated in silica for Enterobacteriaceae. Strains were initially selected based on the presence of the 50 nucleotide SPA fragment Ent2 (see table 36), generated upstream of the RpoBl-R1327 priming site. Subsequently, 50 nucleotide SPA fragments were generated upstream of the RpoB6-R1630 priming site. The sequences of these SPA fragments are presented and for each of these SPA fragments, the Enierobacleriaceae species and the number of strains is indicated. SPA fragments identifying a single strain were left out. Enterohacteriaceae -specific (Ent) SPA fragments received a unique numerical identifier for reference in further analysis with an asterisk symbol indicating that the SPA fragment was generated from the region upstream of the RpoBl-R1630 priming site.

[00250] Comparing the results from Table 36 and Table 37 shows the improved phylogenetic classification of species that clustered together for S PA fragment Ent l after they were further classified using 50 nucleotide SPA fragments generated from the region upstream of the position of the RpoB6-R1630 priming site. For instance, the 1006 Escherichia coli strains previously identified by SP A fragment Bntl broke into several subgroups, most of the 32 Shigella sonnet strains ended up in different groups, as did the 14 Escherichia fergusonii and the 13 Escherichia albertii strains. The results from whole genome-based ANI show that strains identified by SPA fragments Ent3*, Bnt4*, Ent5* and Ent 16*, despite representing different species, are very closely related with ANI scores of >0.97. Members of the genus Shigella have high genomic similarity to Escherichia coli and are often considered to be atypical members of this species. In line with the observation that many Shigella and Escherichia coli strains were identified by the same SPA fragment, Shigella species were reclassified as Escherichia species in the Genome Taxonomy Database (GTDB) using an operational average nucleotide identity (ANI)-based approach nucleated around type strains (Parks et al, 2021).

[00251] SPA fragment Ent7* identified Escherichia coli and Escherichia fergusonii strains, and SPA fragment Ent9* identified Escherichia albertii strains. Based on whole genome-based ANI it can also be concluded thatM/geZ/a boydii strain 60_SBOY (Ent4) should be assigned as Escherichia coli, that Escherichia coli strain 102606_aEPEC- (Ent9) should be reassigned as Escherichia albertii, and that Escherichia coli strain JL_F4J (Entlb) and Shigella sonnei strain ECSW +02 (Ent16) represent the same species with an ANI score of 1.00.

[00252] Similarly, comparing the results from Table 36 and Table 38 shows the improved phylogenetic classification of species that clustered together for SPA fragment Ent2 after they were further classified using 50 nucleotide SP A fragments generated from the region upstream of the position of the RpoB6-R1630 priming site. For instance, SPA fragment Entl 8* specifically grouped closely related Klebsiella pneumoniae, Klebsiella quasipneumoniae and Klebsiella aerogenes strains. This was confirmed by whole genome-based ANI as shown in Figure 32, where two major groups could be distinguished. Strains of ANI group I share very high ANI scores (>99%), indicating that the Klebsiella pneumoniae, Klebsiella quasipneumoniae and Klebsiella aerogenes strains of this group represent members of the same species. Since this group includes the Klebsiella pneumoniae ATCC 43816 type-strain, members of this group should be identified as Klebsiella pneumoniae. Similarly, members of the AN ! group II, which include the Klebsiella quasipneumoniae ATCC 700603 type-strain, should be identified as Klebsiella quasipneumoniae. [00253] In addition to Klebsiella pneumoniae and Klebsiella quasipneumoniae strains, SPA fragment Ent2 identified closely related Enterobacter sp. strains that could be further classified using 50 nucleotide SPA fragments generated from the region upstream of the position of the RpoB6-R1630 priming site, as was confirmed by whole genome-based ANI. Based on the ANI results it can be concluded that many strains that were previously identified as Enterobacter cloacae represent in feet different but closely related species. However, the strains designated as Enterobacter cloacae identified by SPA fragments Ent20* and Ent23* represent true Enterobacter cloacae: this also includes the Enterobacter cloacae ATCC 13047 type-stain. SPA fragment Ent20* also identifies Enterobacter asburiae strains. However, based on their ANI score of 0.88 with Enterobacter cloacae ATCC 13047, the strains identified by SPA fragment Ent24* represent a different species, which is confirmed by their unique SPA fragment.

[00254] SPA fragment Entl9* grouped closely related Enterobacter sp. strains, including Enterobacter fcobei strains, Enterobacter roggenkampii strains, Enterobacter bugandensis stains, and Enterobacter asburiae strains. Based on whole genome ANI, Leclercia adecarboxylata UMB0660 identified by SPA fragment Entl9* represents an Enterobacter bugandensis strain. In addition to SPA fragment Entl9*, Enterobacter asburiae strains were identified by SPA fragment Ent20*, Enf25*, Ent26*, Ent30*, and Ent27*, which also identified the reference strain Enterobacter asburiae 35734 and the type-strain Yokenella regensburgei ATCC 49455

[00255] SPA fragment Ent20* identified strains from the closely related species Enterobacter cloacae and Enterobacter asburiae. Serratia fonticola strains were specifically identified by SPA fragments Ent22* and Ent31*. SPA fragment Ent28* was found to be specific for Enterobacter mori, while SPA fragments Ent21 * and Ent29* were found to be specific for Leclercia adecarboxylata and a closely related Leclercia species; this species was also identified by SPA fragment Ent25*. The results also show that Leclercia adecarboxylata strain UMB0660, identified by SPA fragment Ent 19*, should be reassigned to Enterobacter bugandensis. The results for the Enterobacter iaceae specific SPA fragments are summarized in Table 39.

Table 39: Summary of Emerobacieriaceae species (Ent) specific SPA fragments as phylogenetic identifiers at the species level. The 50 nucleotide SPA fragments are identified as SPA fragment “Ent” and a numerical identifier, with an asterisk symbol indicating that the SPA fragment was generated from the region upstream of the RpoBl-R1630 priming site.

[00256] To show the synergy of using two SPA fragments generated from two distinct regions of the rpoB gene for phylogenetic identification of closely related bacteria we compared the phylogenetic classification of 121 Escherichia coll strains and related species belonging to different phylotypes as described by Fang et al (2018). 'Phis includes Escherichia coli phylotype B2 strains, which are prevalent in IBD patients and have distinct metabolic capabilities that allow them to colonize mucosa. The results are presented in Figures 33A, 33B and 33C. Figure 33A shows the phylogenetic free of the strains when the sequences of 50 nucleotide SPA fragments generated from the region upstream of the RpoB 1-R 1327 priming site are used. Except for a subset of Escherichia coll phylotype B2 strains and a small group of Escherichia coll phylotype B2 and D strains, all strains clustered together, including the Shigella species that are closely related to Escherichia coli phylotype A and Bl strains. Figure 33 B shows the phylogenetic tree of the strains when the sequences of 50 nucleotide SPA fragments generated from the region upstream of the RpoB 1~R 1630 priming site are used. This resulted in a significant improvement of the phylogenetic clustering, especially for the Escherichia coli phylotype B2 strains. Figure 33C shows the phylogenetic tree of the strains when the combination of sequences of 50 nucleotide SPA fragments generated from the regions upstream of the RpoB 1-R 1327 and RpoB6-R1630 priming sites are used. The combined use of SPA fragments that represents different gene regions with phylogenetic information refines the phylogenetic clustering of the Escherichia coli strains, including the phylotype B2 strains, to a resolution that is not obtained when any of the two fragments are individually used. Therefore, for the identification of closely related species, the SPA fragment method (Figure 2) can include one or more additional primers to simultaneously target different regions for phylogenetic identification. These regions can be located on the same gene, as demonstrated for the rpoB gene, or on different phylogenetic genes, especially conserved housekeeping genes. Subsequently, data from the individual primers are processed for community composition and species identification. In case of inconclusive identification, the information from both SPA fragment sets is combined to enhance the phylogenetic resolution. In addition, having more than one primer serves as an internal control for community composition. Overall, the results demonstrate how the disclosed SPA fragment sequencing method is generalizable and adaptable to improve phylogenetic resolution in a targetable fashion for the identification of closely related species of clinical importance, including members of the Enterobacteriaceae.

[00257] In certain instances, disease phenotypes caused by bacteria will depend on the presence of virulence/pathogenicity factors located on mobile genetic elements, including conjugate ve and/or mobile plasmids, phages, and pathogenicity islands that can be horizontally transferred between bacteria, as is the case for Escherichia coll, Salmonella, Klebsiella, Listeria, Bacillus, pyogenic streptococci and Clostriclium perfringens, among others (for review, see Gyles and Boerlin, 2014). As the result of horizontal gene transfer, phylogenetic information on species composition will be insufficient to predict disease pathology, and therefore needs to be complemented with information on community functionality. For instance, the presence in Escherichia cali of the PKS pathogenicity island encoding, among other virulence factors, for genotoxic colibactm synthesis has been linked to increased risk for developing colorectal cancer (Pleguezuelos-Manzano et al, 2020). As discussed for colorectal cancer in Example 7, multiplex SPA fragment sequencing provides the flexibility to address both phylogenetic identification and community functional in a same amplification step. By designing a primer for SPA fragment amplification that specifically targets the PKS gene cluster essential for colibaciin synthesis, the presence of genotoxic Escherichia coll strains (Pleguezuelos-Manzano et al, 2020) can be determined, and combined with phylogenetic information be used for improved risk assessment and detection of colorectal cancer.

EXAMPLE 10: SENSITIVITY ANALYSIS OF SPA FRAGMENT SEQUENCING

[00258] understand the sensitivity of the SPA fragment sequencing method, the gut microbiome community of a person suffering from intestinal complications was used for in silica simulations. The assumption was that this microbiome would leave a similar signature in the tncfDNA. This consortium (see Table 40), whose composition was determined using long-read PacBio sequencing, is interesting as it includes six Metagenome Assembled Genomes (MAGs) representing closely related Faecalibacleriutn strains that based on their Average Nucleotide Identity (ANI) represent five different species- subspecies (Figure 34A). Therefore, one of the questions is whether the SPA fragment sequencing method can provide the level of phylogenetic resolution to discriminate between these strains, and if this would be at the 25 base pair or 50 base pair SPA fragment length. This consortium also includes three MAGs representing Bacteroides ovatis, which were found to be very similar based on their AN! score of 0.99 (Figure 34B), and that their assignment to different MAGs was most likely the result of binning errors. As such it is expected that these strains would share the same SPA fragment. Since the PacBio sequencing did not result in complete MAGs for all strains, especially for strains with lower abundances, whole genome sequences from the closest related strains as identified with ANI were used in the simulations.

Table 40: Composition (species name and genome ID) and relative species abundances of the gut microbiome community used for the simulations. Strains with identical SPA fragments of

25 base pairs (see Table 41 ) are indicated by the same *number. Relative Abundance %

(number of genome copies of each species Z sum of genome copies of all species) x 100%.

Fnmfr/bflczt’ntjffl species are in marked in bold. members: To demonstrate the discriminatory power of SPA fragment sequencing targeting the RpoB gene, 25 base pair and 50 base pair long SPA fragments located 3* of the RpoB 1-R 1327 primer annealing site were generated in silica for each of the community members. The results for the 25 base pair long SPA fragments that identified more than one bacterial strain present in the community are presented in Table 41. Identical results were obtained for the 50 base pair SPA fragments. It should be noted that for the simulations, we still consider that all strains can be identified by their individual SPA fragments.

[00260] Using the sequences of either the 25 or 50 base pair long SPA fragments, 50 of the 52 strains in the community could be identified on the species level by their unique SPA fragments. Four SPA fragments obtained in silico with the RpoBl-R1327 primer identified multiple bal very closely related strains (Table 41), as was confirmed by their identical genome taxonomy. Based on genome taxonomy and AN I it was concluded that each recognized strains belonging to the same species, and that their assignment to different MAGs was most likely the result of binning errors.

[00261 ] The six Faecalibacterium strains, classified on whole genome-based AN I as belonging to five different (sub)species ( Figure 34A), were each identifiable by their unique SPA fragment sequence of 25 base pairs or longer, except for two strains that both belonged to Faecalibacterium prausmteii subgroup G, and that shared AN! scores of 97%, indicating that they represent the same species, as confirmed by these two strains sharing the same 50 base pair long SPA fragment. As such whole genome-based AN! and SPA fragment sequences provided the same phylogenetic resolution to discriminate these strains at the (sub)species level. The Baeieroides ovalus strains, that based on genome taxonomy and whole genome ANI were closely related and represented the same species (Figure 34B), shared the same 25 base pair and 50 base pair SPA fragment sequence, also pointing to similar phylogenetic resolution of the two methods. The only exception was for the two closely related Roseburia species that shared common 25 and 50 base pair long SPA fragments, but that according to their genome taxonomy based on the Genome Taxonomy Database (Parks et al, 2018) represented two different species, Overall, these results confirm the specificity of SPA fragment sequences obtained 3 ’ of the RpoR1-Rl327 primer annealing site for the high-resolution identification of bacterial strains at the (sub)species level.

Table 41 : Overview of 25 base pair long SPA fragments with more than one identified bacterial strain in the consortium. The detailed genome taxonomy is based on the Genome Taxonomy Database (Parks et al, 2018). The nucleotide sequences of the 25 base pair long SPA fragments are included. d_: domain: p_: phylum; c_: class; o_: order; f_: family; g_: genus; s_: species.

[00262 ] Description of the parameters to simulate the effect of SPA fragment length on community comppsiiion : Four simulations, each having 30 trials, were run with varying average length of mcfDNA fragments (40, 60, 80 and 100 base pairs). For the simulations we used of 1 ml liquid biopsy sample containing 100ng/ml cfDNA and assumed that 1% of the total cfDNA represents mcfDNA (Ing/ml). These estimates are considered realistic; for instance, in patients with metastatic breast cancer, the median plasma cfDNA concentration was found to be 112 ng/inl (Fernandez-Garcia el al, 2019). To be very conservative, we also estimate that due to technical limitations only 10% of the mcfDNA is effectively processed. As such, the simulations assume that fragments are only generated from 0.1 ng mcfDNA.

[00263] For each genome in the microbial community, length weighted relative abundance of total sample fragments was determined to account for the larger number of mcfDNA fragments generated from larger bacterial genomes. This abundance was subsequently used to determine the number of mcfDNA fragments per genome. The mcfDNA fragment sizes are randomly selected using a truncated normal distribution with fragment sizes between 1 and 200 base pairs. The fragment ends (start and end positions) were randomly selected from the genome. If a fragment contains the SPA primer annealing site, an in silico SPA fragment is generated from the 3’-end of the SPA primer to the end of the fragment (Figure

[00264] As described herein above, SPA fragments of 50 base pairs or longer, obtained using the RpoBl-R1327 primer, provide high resolution phylogenetic identification for most bacteria at the species and subspecies level. Therefore, the “number of SPA fragments generated with length 50 base pairs or greater” is used as one of the criteria to determine the sensitivity of the method for species identification in function of the various parameters. It should also be noted that many more SPA fragments with smaller length will be generated.

[00265] As previously shown herein above SP A fragments with length 25 base pairs or greater, obtained using the RpoBl-R1327 primer, show good resolution at the genus level. Therefore, the “relative abundance numbers of SPA fragment with length 25 base pairs or greater” will be used to calculate the community composition.

[00266] The parameters used in the 'four simulations are presented in Table 42. The following formula is used to calculate the “total number of cfDNA molecules”, based on X ng cfDNA with an average length of ¥ bp for the mcfDNA: (X ng x [6.022* 10 ²³ ] molecules/rnol) / (Y bp ^x [1 ^x 10 ⁹ ]ng/g ^x 618 g/mol).

Table 42: Overview of the conditions used for the simulations to determine the sensitivity of the SPA fragment sequencing method. The estimate of generated mcfDN A fragments being 0.1% of the cfDNA is based on the conservative assumption that 1% of cfDNA represents mcfDNA, and that due to technical limitations and losses during processing steps, approximately 10% of mcfDNA fragments will be correctly processed and contribute to SPA fragments.

[00267] Simulation of fragment size distributions: We first evaluated the distribution of fragment sizes. To do so, we simulated the size distribution of a million mcfDNA fragments based on a truncated normal distribution with averages of 40, 60, 80 and 100 nucleotides in length, respectively. The results are presented in Figure 35. Of the four simulations, the size distribution obtained for the simulation around an average fragment length of 60 base pairs came closest to the reported size distribution for mcfDNA (Burnham et al, 2016). We therefore consider this simulation the most relevant. The simulation for fragments with an average length of 40 base pairs missed nearly all fragments larger than 70 base pairs, while the simulations for fragments with average lengths of 80 base pairs and 100 base pairs underrepresented the smaller fragments and overrepresented fragments larger than 100 base pairs.

[00268 ] Simulation of SPA fragment generation for species identification and community composition analysis : For each simulation, the trial was repeated 30 times. The Wilcoxon rank sum test was performed on each of the simulations, by genome, with the two null hypotheses being: “the count of SPA fragments of 50 base pairs or greater was less than 3” (key criterium for species identification); or “the count of SPA fragments of 25 base pairs or. greater ..was Jess ..than.. 10” (key criterium for species abundance). The results for the simulations using mcfDNA fragments with an average length of 40 base pairs or 60 base pairs are presented in Table 43 and Table 44, respectively; the R.poB1-R1327 was used to create the SPA fragments targeting the rpoB gene for phylogenetic identification.

[00269] Based on the results presented in Table 43, the null hypotheses “the count of 3 SPA fragments of 50 base pairs or greater was less than o” gets accepted for the simulation using mcfDNA fragments with an average length of 40 base pairs. This indicates that for the conditions used in this simulation no reliable strain identification can be obtained at. the species and subspecies level based on the presence of SPA fragments of 50 base pairs or greater. However, the null hypothesis “the count of 10 SPA fragments of 25 base pairs or greater was less than 10” gets rejected for strains that are present at approximately 1.25% or above with a p-value <0.05. This indicates that under the simulated conditions, using mcfDNA with an average fragment length of 40 base pairs, species present at approximately 1 .25% in the community can be reliably identified by their 25 base pairs or greater SPA fragments at the genus level, and in many cases at the species level, in addition, the relative abundances of these species can be calculated.

[00270] Based on the results presented in Table 44, the null hypotheses “the count of SPA fragments of 50 base pairs or greater was less than 3" (key criterium for species identification) and “the .count of S PA_ fragments of 25 base .Jiairs or greater

(key criterium for species abundance) both get rejected with a p-value <0.0001, This indicates that mcfDNA fragments with an average length of 60 base pairs can be reliably used for the identification of strains at the species and subspecies level, when the strains represent approximately 0.5% of the microbial community composition. In addition, mcfDNA fragments with an average length of 60 base pairs can be used to determine the community composition for species present at approximately 0.5%. Very similar results were obtained for the simulations using average mcfDNA fragment lengths of 80 base pairs and 100 base pairs.

[00271] On average, approximately 14,500 mcfDNA fragments that contain the RpoB1-R1327 primer annealing site were generated per trial for the simulation using mcfDNA fragments with an average length of 60 base pairs, of which approximately 5650 fragments would generate SPA fragments of 25 base pairs or greater. This should provide ample targets for the amplification step in the SPA fragment protocol, and subsequent sequencing,

[00272] Conclusions: Overall, the simulations show that mcfDNA fragments with an average length of 60 base pairs can be reliably used for the identification of strains at the species and subspecies level when they are present at 0.5% or above in the microbial community detectable in liquid biopsy samples, including peripheral blood. On average, strain abundances measured based on SPA fragments were within 1.4% of the actual abundance. For strains with less than 1% abundance, the average error was 1.8%, ranging from 0.1% to 7.2%.: for strains with an abundance of 1% or higher, the average error was 1 .2%, ranging from <0.1% to 4.5%.

Table 43: Summary of Simulation 40-lOOng (average generated nicfDNA length of 40, lOOng of cfDNA) using the RpoB1-R1327 primer. Bacterial sped es, represented by their genome ID, whose presence and abundance were considered as significant (p-value< 0.05) are highlighted in grey. Total mctDNA Fragments per Genome with Conserved Region for Primer indicates the total number of fragments generated for the 30 trials of the simulation. SPA Fragments >

24 bp long refers to SPA fragments of 25 base pairs or greater; SPA Fragments > 49 bp long refers to SPA fragments of 50 base pairs or greater.

Table 44: Summary of Simulation 60-100ng (average generated mcfDNA length of 60. lOOng of cfDNA) using the RpoBl-RI 327 primer. Bacterial species, represented by their genome ID, whose presence and abundance were considered as significant (p-value< 0.05) are highlighted in grey. Total mcfDNA Fragments per Genome with Conserved Region for Primer indicates the total number of fragments generated for the 30 trials of the simulation. SPA Fragments -> 24 bp long refers to SPA fragments of 25 base pairs or greater; SPA Fragments > 49 bp long refers to SPA fragments of 50 base pairs or greater.

EXAMPLE 1 .1 : SPECIFICITY ANALYSIS OF SPA FRAGMENT SEQUENCING

[00273] Several studies have shown that high resolution phylogenetic identification of bacteria is a prerequisite to accurately link bacteria to specific disease phenotypes, including the development of adenomas and eariy-stage carcinomas in colorectal cancer. Therefore, one of the key requirements for SPA fragment sequencing is high-resolution identification of microbial species in liquid biopsy samples at the species and subspecies level . We therefore tried to answer the following questions: 1. Specificity of SPA fragments - How phylogcnetically accurate arc the 46 distinct 50 base pair long SPA fragments generated using the RpoBI -R 1327 primer (EXAMPLE 11)?

2. How does the sensitivity and specificity of the SPA fragment sequencing method compare to deep rnetagenome sequencing of cfDNA fragments followed by taxonomic classification using read-based metagenome analysis methods ( EXAMPLE 12)?

[00274] Description of the community usc4 for Jh& ^simulations^ To understand the specificity of the SPA fragment sequencing method, the same gut community described in EXAMPLE 10 was used for the simulations. The 52-tnembcr community, whose composition was obtained with PacBio sequencing, is described in Table 45. The sequences of the SPA fragments obtained for each of the comm unity members are also presented. SPA fragments that were identical between multiple community members are highlighted in grey.

Table 45: Composition (species name and genome ID) and relative species abundances of the gut microbiome community used for the simulations. Long read PacBio sequencing was used to determine the community composition. The community composition based on the rpoB gene-derived SPA fragment sequencing simulation was determined using the parameters described above. The codes and sequences for the unique 50 base pair SPA fragments generated for each species are shown. SPA fragments that arc identical between multiple community members are highlighted in in grey,

[00275] Specificty analysis of SPA Fragments obtained using the RpoB1-R1327 primer: To analyze the phylogenetic specificity of the SPA fragments listed in Table 45. we compared them to a phylogenetic gene database containing over 50,000 unique RpoB gene entries. The results of this comparison are presented in Table 46 and show the following:

• The 50 base pair SPA fragments for the 52 community members showed 100% correct phylogenetic identification on the genus level and were also highly specific on the species level when compared to the reference database of 50,000+ non-redundant RpaB gene entries. Three of the SPA fragments identified multiple, closely related species: o In addition to recognizing Baclemides ovatus, the SPA2 fragment also recognized the closely related species Bacteroides xylanisolvenst, and in addition to recognizing Alistipes onderdonkii, the rpob_SPA46 fragment also recognized the closely related species Alistipes finegoldii and Alistipes shahii. o The rpob_SPA8 fragment recognized the Blautia_A wexlerae_A, Blautia_A wexlerae and BIautia_A sp003480185, which according to the new' classification of the Genome Taxonomy Database (Parks et al, 2018) represent very closely related but distinct species; the same is the case for the rpob_SPA40 fragment, which recognizes the very closely related but distinct species Roseburia inulinivorans and Roseburia sp900552665.

• The fragment tpob_SPA21 enabled identification of Bifidobacterium longum at the species level but failed to discriminate on the subspecies level between Bifidobacterium longum subsp. longum and Bifidobacterium longum subsp. infantis; and the rpob_SPA24 fragment enabled identification of Anaerostipes hadrus at the species level but failed to discriminate on the subspecies level between Anaerostipes hadrus and Anaerostipes hadrus _B.

• It was also noted that the Faecal i bacterium species present in the community could be identified to the species level by their unique SPA fragment, and in several cases to the Faecalibaeterium prausnitzii subspecies level. The only exception was the fragment rpob_SPA18, which recognized the two very closely related subspecies .Faecalibaeterium prausniteii J and Faecalibaeterium prausnitzii.

Overall, the results shows that SPA fragments generated 3’ of the RpoBl~R1327 primer annealing site have very high phylogenetic specificity to reliably classify bacteria at both the taxonomic genus and species level.

Table 46: Simulated composition of the gut microbiome community based on rpoB gene- derived SPA fragment analysis. Each community member is identified by its GTDB taxonomy and PATRIC genome ID. The genus-level and species-level identification of each community member, based on its 50 base pair rpoB gene-derived SPA fragment, is presented based on GTDB taxonomy (Parks et al, 2018). For each community member, the relative abundance and

SPA fragment identifier are listed. SPA fragments, which identified multiple species, are highlighted in grey.

EXAMPLE 12: SIM ULATION OF SENSITIVITY AND SPECIFICITY ANALYSIS OF DEEP NEXT GENERATION SEQUENCING

[00276] sequencing of mcfDNA fragments followed bv taxonomic classification using read-based melagenome analysis methods: The current approach to analyze microbial signatures in cfDNA involves deep NGS sequencing. After filtering out the human DNA reads, the mcfDNA reads are analyzed; this is customary done using read-based taxonomic classifiers. To understand the usefulness of read-based taxonomic classifiers for mcfDNA informed community analysis we simulated mcfDN A fragments and classified them with either Kai ja (Menzel et al, 2016) or Kraken 2 (Wood et al, 2019), two commonly used read-based taxonomic classifiers. For this simulation we used the assumption that on a routine basis 100 cfDNA samples were sequenced in parallel on a NovaSeq 6000 NGS sequencer. Since the maximum capacity of the NovaSeq 6000 is approximately 20 billion reads, this would enable sequencing of a maximum of 200 million cfDNA fragments per sample, This is in line with the numbers published by Poore et al (2020). Based on the assumption that 1% of the cfDNA represents mcfDNA fragments, around 2 million mcfDNA fragments sequence reads will be generated per sample.

[00277] For each genome in the microbial community of T able 40, the length weighted relative abundance of total sample fragments was determined to account for the larger number of mcfDNA fragments generated from larger genomes. This abundance was subsequently used to determine the number of mcfDNA fragments generated per genome. The mcfDN A fragment sizes were randomly selected from a truncated normal distribution with fragment sizes between 1 and 200 base pairs and an average of 60 base pairs; these represents the same parameters as used for the SPA fragment simulation and matches best with the reported size distribution for mcfDNA fragments (Burnham et al, 2016). The fragment start and end positions were randomly' selected from the genomes.

[00278 ] The results of the taxonomic assignment of fragments by Kaiju and Kraken 2 to different phylogenetic levels, ranging from phylum to species, is presented in Table 47. The community compositions determined by PacBio sequencing and the SPA fragment sequencing simulation using the RpoBl-R1327 primer are included for reference. Based on the results presented in Table 47 it can be concluded that Kaiju and Kraken 2 failed to correctly assign short mcfDNA reads to their taxonomic classification or to correctly deconvolute the community composition. This is in contrast to the results obtained for the SPA fragment sequencing simulation, which closely matched the community composition obtained by PacBio sequencing that was used as input for all three simulations. It is also important to remember that for all three simulations, similar mcfDNA fragments with an average length of 60 base pairs and a similar size distribution were used.

Table 47: High-level phylogenetic breakdown and assignment of simulated mcfDNA reads to different phylogenetic levels by Kaiju and Kraken 2. For comparison, phylogenetic breakdown of the community obtained by PacBio sequencing and simulated SPA fragment sequencing are included. The numbers between brackets represent the number of reads that were assigned by Kaiju and Kraken 2 to a phylogenetic level; this excludes fragments identified as viruses and unclassified reads.

[00279] Further details on the phylogenetic assignment of mcfDNA reads to the genus level by Kaiju and Kraken 2 are presented in Table 48 and Table 49, respectively. In the original community, all 52 members are present at a relative abundance ranging from 0.54% to 21.61% (see Table 40). Of the reads, 40.77P4» and 38.04% could be assigned by Kaiju and Kraken 2, respectively, to the genus level, represented by genera with a relative abundance of 0.01% or above. This number is in line with the results published by Poore et al (2020), with 35.8% of the mcfDNA reads being assigned to the genus level. A further comparison of the genus level taxonomic assignment is provided in Table 50.

Table 48: Composition on the genus level of the simulated gut microbiome community using Kaiju (version 1.7.2) for taxonomic classification of in silico generated mcfDNA fragments.

Table 49: Composition on the genus level of the simulated gut microbiome community using

Kraken 2 (version 2.08) for taxonomic classification of in silico generated mcfDNA fragments.

Kraken 2 (version 2.1.2) was also run with no significant improvement in the results.

Table 50: Comparison between the composition on the genus level of the gut microbiome community between the SPA fragment sequencing simulation and simulated NGS sequencing of mcfDNA using Kaiju or Kraken 2 for taxonomic classification. To facilitate comparison, some of the genera listed in Table 46 have been combined, reducing the total number of genera from 27 to 25. N.A.: not applicable; the genus was either not found or no reads were assigned to it. The genera Phocaeicola and Mediterraneibacter were not present in the databases used for taxonomic classification by Kaiju or Kraken 2, and their abundances were included in the genera Bacieroides and Ruminococcus^ respectively, to which they previously belonged,

[00280] Based on the results presented in Table 50 it can be concluded that all three simulations identified the most abundant genera, including Bacteroicles, Blautia, Faecalibacterium, Alislipes, Phocaeicola, Agalhobactduoi/ParaprevoleUa, Bifidobacterium and Fusicatenibacter. However, compared to the input data for the simulations, the numbers for their relative abundances are imprecise for Kaiju and Kraken 2. This becomes even more obvious for low abundant species. In addition, the read-based taxonomic classification tools fail to provide any meaningful insights when multiple closely related species are present.

[00281] Species and subspecies level insights are required to draw meaningful conclusions between microbial signatures and diseases, including cancer detection and prognostics. The simulated compositions on the species level of the gut microbiome community using Kaiju or Kraken 2 for taxonomic classification of in silico generated mcfDNA fragments were very imprecise. For Bacteroides stercoris, the dominant species present at 21 .61 % in the community, Kaiju was able to match 2.6% of the mcfDNA fragments to this species, while Kraken 2 failed to link any mcfDNA fragments to this species. This clearly shows that read-based taxonomic classification tools are lacking the sensitivity and specificity required to analyze microbial signatures present in mcfDNA from biopsy samples.

[00282] Conclusion: Short DNA fragments with an average length of approximately 60 base pairs are an intrinsic property of mcfDNA. In contrast to the result from the simulation using SPA fragment sequencing-based analysis, where the fragments were generated using the RpoBl-R1327 primer, simulations using deep metagenome sequencing of cfDNA fragments followed by taxonomic classification of mcfDNA using read-based metagenome analysis methods showed that the current read-based tools are unsuitable for taxonomic classification of the short sequencing reads obtained from mcfDNA. As such this approach lacks the sensitivity and specificity to provide meaningful insights for disease detection and progression monitoring. An approach to overcome this limitation would require very deep sequencing and assembly of short reads into larger fragments. In addition to a significantly higher sequencing cost, limitations in the assembly of short sequencing reads makes this approach unsuitable for scalable application to the routine analysis of microbial patterns in biopsy samples.

EXAMPLE 13: CPN60 GENE-BASED SPA FRAGMENT SEQUENCING [00283] As concluded from EXAMPLE 11, SPA fragment sequences obtained with the primer RpoBl-R1327 provided excellent phylogenetic resolution for gut microbiome bacteria at the genus level and in most instances at the species and subspecies level. However, in some instances, it failed to discriminate between very closely related species, such as Bacleroides ovalus and Bacteroides xylanisolvens., and Alistipes cmderdonkU* Alistipes fmegoldii and Ah'stipes shahii.

[00284] Design of the Cpn60~E571 SPA primer: To further improve the phylogenetic resolution compared to SPA fragment sequencing based on the rpcB gene (using primer RpoBl -RI327) we analyzed the 60 kDa chaperon in protein gene (cpn60 gene, also referred to as the groEL gene) for SPA fragment sequencing. Using the method described herein above and exemplified in Example 2, a conserved region spanning position 571 to 593 (position numbers based on the Escherichia coli cpn60 gene) was identified for SPA fragment sequencing; this primer annealing region is located downstream of a hypervariable DNA region to be used for phylogenetic identification. The degenerate nucleotide sequence of this region is presented in Figure 7B. The primer Cpn60-R571 was tested for SPA fragment amplification of the region upstream of position 571 of the cpn60 gene as described in this Example. The Cpn60-R571 primer has the sequence listed below, using the following nucleotide codes: A: adenine; G: guanidine; C: cytosine: T: thymine; R: purine (A or G); Y: pyrimidine (T or C); K: amino (T or G); B: not A (T, G or C); N: any nucleotide (A, G, C or T).

Cpn60-R571 primer: 5’ CCN.YKR.TCR.AAB.YGC.ATN.CCY.TC 3’

[00285] As described herein above, a conserved primer annealing region is located adjacent to a t least one of a 25 nucleotide-long or a 50 nucleotide-long variable region with preferably an average sequence variance of <0.1 and <0.075, respectively. As can be seen in Table 51 , the 25 nucleotide-long variable region located upstream of the Cpn60-R571 primer annealing site has an average sequence variance of 0.0851. Tabic 51: Average sequence variance for the Cpn60-R571 primer region and the regions upstream or downstream of the primer annealing region. For both regions located adjacent to the primer region, the variance is shown for 25, 50, 75, 100 or 200 nucleotides (nt) upstream (5’) or downstream (3’) of the beginning or end of the primer annealing sequence. The variance score is calculated as the average of the variance of the percentage of the nucleotides adenine, guanidine, cytosine and thymine at each position of the cpn60 gene. A lower number is indicative for more variance, while a higher number is indicative lor less variance and a more conserved DNA sequence. The maximum theoretical variance score for a region is 0.25 (would represent a 100% conserved DNA region). Regions with a variance score <0,1 are highlighted in grey. sequences: Using a similar consortium (see Table 40) and parameters for the simulations as described in EXAMPLE 10, a simulation was performed to determine the sensitivity of SPA fragment sequencing using the Cpri60-R571 primer annealing site. I'he 52-member community, whose composition was obtained with PacBio sequencing, is described in Table 52. The sequences of the SPA fragments obtained for each of the community members are also presented. The 50 base pair SPA fragments that are identical for multiple closely related community members are highlighted in grey. Based on the results from EXAMPLE 10, mcfDNA fragments with an average sequence length of 60 base pairs were used in this simulation. The results from the simulation using the Cpn60-R571 primer showed that. mcfDNA fragments with an average length of 60 base pairs can be reliably used to determine the microbial community composition when the strains are present at approximately 0.5% (Table 53). These results are very similar to the results that were obtained for the simulation using the RpoB 1-R1327 primer (Table 44).

Tabic 52: Composition (species name and genome ID) and relative species abundances of the gut microbiome community used for the simulations. Long read PacBio sequencing was used to determine the community composition. The community composition based on the SPA fragment sequencing simulation was determined using the parameters described above and is also presented in Table 53. The codes and sequences for the unique 50 base pair SPA fragments generated for each species are shown. SPA fragments that are identical for multiple community members are highlighted in in grey. Compared to the strain selection for the RpoBl -RI327 simulation, two strains for which no cpn60 gene could be identified were replaced by closely related strains: Faecalibacterium prausnilzii strain COPD342 and Ruminococcus sp. CAG:9 were replaced by Faecalibacterium prausnilzii strain S()3C.ntcta.bin_9 and Blatitia wexlerae strain S09A.meta.bin 3, respectively.

Table 53: Summary of Simulation 60-100ng (average generated mcfDNA length of 60, 100ug of cfDN A) using the Cpn60-R57l primer. Bacterial species, represented by their genome ID, whose presence and abundance were considered as significant (p-value< 0.05) are highlighted in grey. Total mcfDNA Fragments per Genome with Conserved Region for Primer indicates the total number of fragments generated for the 30 trials of the simulation. SPA Fragments > 24 bp long refers to SPA fragments of 25 base pairs or greater; SPA Fragments > 49 bp long refers to SPA fragments of 50 base pairs or greater.

[00287] Specificity analysis of SPA fragments obtained using the Cpn60~R57I primer: To analyze the phylogenetic specificity of the CpnbO-SPA fragments listed in Table 52, we compared them to the reference phylogenetic gene database which contains over 40,000 unique cpnfiO gene entries. We also compared the results with those obtained using the RpoBI -R1327 primer. The results of this comparison are presented in Table 54 and show the following:

• As shown in EXAMPLE 1 1 , the rpoB gene-derived SPA fragments rpob_SPA4, rpob_SPA8, rpob_SPA40 and rpob_SPA46 were unable to distinguish between closely related Bacieroides, Blautia_A, Roseburia and Alistipes species, respectively; and SPA fragments rpob_SPA18, rpob_SPA2l and rpoB_SPA24 failed to discriminate on the subspecies level between Faecalibacterium prausnitzii and Faecal ibacterium prausnitzii J, Bifidobacterium longum subsp. kmgum and Bifidobacterium longum subsp. infantis, and Anaerostipes hadrus and Anaerostipes hadrus_B, respectively,

• The cpn60 gene-derived SPA fragments cpn60_SPA5 and cpn60_SPA8, cpn60_SPA7 and cpnbO _SPA.37, cpn60_SPA19, cpn6() _SPA34 and cpn60 JSPA4O were unable to distinguish between closely related Blatdia_A, Faecalibacterium, Acetatifactor. Bacteroides and Roseburia species, respectively; and SPA fragment cpn60_SPA24 and SPA fragment cpn60_SPA34 failed to discriminate on the subspecies level between Anaerostipes hadrus and Anaerostipes hadrusJB, and Bifidobacterium longum subsp. longum, Bifidobacterium longum subsp. infan tis and .Bifidobacterium longum subsp. imperatoris, respecti vely.

• It should also be noted that the simulated community compositions using rpoB gene- derived SPA fragments and cpn60 gene-derived SPA fragments are very similar.

[00288] Unexpectedly, the phylogenetic resolution on the species level was gene dependent and, therefore, combining the results from multiple phylogenetic genes will result in better phylogenetic deconvolution of the community. As shown in Table 54, in several cases where SPA fragments derived from a single phylogenetic identifier gene failed to provide species level resolution, the combination of rpoB and cpn-60 gene-derived SPA fragments from the same species allowed for improved phylogenetic resolution at the species level. Improved phylogenetic identification on the species level by rpoB gene-derived SPA fragments (compared to cpnbO gene-derived SPA fragments) was observed for Blautia_A massiliensis (rpoB SPA5 fragment), Faecalibacterium prausnitzii C (rpoB SPA 7 fragment), Blautia A sp(103480185 (rpoB_SPA12 fragment), Acetatifactor sp900066565 (rpoB_SPA20 fragment), Bacteroides caccae (rpoB SPA35 fragment), and Faecalibacterium prausnitzii JD (rpoB_SPA38 fragment); and improved phylogenetic identification on the species level by cpn60 gene-derived SPA fragments (compared to rpoB gene-derived SPA fragments) was observed for Bacteroides ovatus (cpn60_SPA4 fragment), Roseburia sp900552665 (cpn60_SPA39 fragment) and Alistipes onderdonkii (cpn60_SPA46 fragment); and on the subspecies level for Faecalibacterium prausnitz_J(cpn60__SPA17 fragment). Thus, using the combination of rpoB and cpn60 gene-derived SPA fragments, species-level taxonomic classification ambiguities were solved for Faecalibacterium, Acetatifaclor and Bacteroides, and remained for Blauiia_A species (rpob_SPA8 and cpn60_SPA8 fragments) and Roseburia species (rpob_SPA40 and cpn60_SPA40 fragments); and subspecies-level taxonomic classification ambiguities were solved for Faecalibaclerium prausnitzii and remained for Bifidobacterium kmgum (rpob SPA2 i and cpn60 SPA20 fragments) and Anaerostipes hadrus (rpob_SPA24 and cpn60__SPA23 fragments).

[00289] Based on this result a new method is provided, referred to as multi loci SPA fragment sequencing, which combines SPA fragments from multiple phylogenetic identifier genes to analyze the composition of microbial communities as is described in EXAMPLE 14

fable 54: Simulated composition of the gut microbiomc community based on rpoB and cpn60 gene-derived SPA fragment analysis. Each community member is identified by its GTDB taxonomy (Parks et al, 2018). The genus-level and species-level identification of each community member, based on 50 base pair long rpoB and cpn60 gene-derived SPA fragments, is also presented based on their GTDB taxonomy. For each community member the relative abundances and SPA fragment identifiers are listed. SPA fragments, which identified multiple community members, are highlighted in grey. In case the rpoB and cpn60 gene-derived SPA fragments provided different levels of phylogenetic resolution, the SPA fragment identifier that provided the best phylogenetic resolution and its corresponding species are highlighted in bold.

EXAMPLE 14: MULTI LOCI SPA FRAGMENT SEQUENCING FURTHER IMPROVES SPECIFICITY

[00290] As concluded from EXAMPLE 11 and EXAMPLE 13, SPA fragment sequences obtained with the primers RpoB 1-R.l 327 and Cpn60-R571 provided excellent phylogenetic resolution for gut microbiome bacteria at the genus level and in many instances at the species and subspecies level. However, in some instances, these SPA fragments failed to discriminate between very closely related species and subspecies. To further improve the phylogenetic resolution of SPA fragment sequencing we provide a new approach, referred to as “.Multi Loci SPA Fragment Sequencing”; hi this approach two or more phylogenetic identifier genes are targeted using different gene-specific SPA primers in the same amplification reaction via multiplexing PUR . One example of a protocol is as follows:

• Isolation of cfDNA using standard protocols.

• End repair and 5 ’-phosphorylation of cfDNA fragments followed by the 3’ addition of a deoxy-adenine to create a 3 ’-sticky end formed by a single adenine nucleotide using standard protocols.

• Ligation of an adaptor, which in this embodiment is an asymmetric linker cassette created by annealing the primers SPA-casl and SPA-cas2, using T4 DNA ligase.

• Single point linker cassette repair. To generate multi loci SPA fragments, multiplexing PCR is performed on the ligation product using three primers: (a) the SPAl -amp primer that recognizes the repaired 5’ asymmetrical end of the linker cassette; (b) a primer that recognizes the primer annealing site specific lor the conserved region of the first phylogenetic marker gene, such as the RpoBl-R1327 primer; and (c) a primer that recognizes the primer annealing site specific for the conserved region of the second phylogenetic marker gene, such as the Cpn60-R57 l primer. All primer sequences are provided in Table 1.

• Once the asymmetric linker cassette has been repaired, the primer (SPA 1-amp primer) that recognizes the repaired 5’ asymmetrical end of the linker cassette can anneal and PCR amplification is initiated. In the case of the reverse RpoB1-R1327 and Cpn60-R571 primers, this will result in the amplification of DMA sequences located upstream of position 1327 of the rpoB gene and upstream of position 57.1 of the cpn60 gene, respectively.

• In a follow up PCR step, adapter sequences are added to the amplified SPA fragments using the primers RpoB1-SPA-seq-R1327, Cpn60-SPA-seq-R571 and S1-PA1--seq-F (see Table 1 ). Alternatively, these primers can be directly used in STEP 4. Subsequently, multiplexing indices and sequencing adapters, such as Illumina sequencing adapters, can be attached using the Nextera XT Index Kit, after which fragments are paired-end sequenced using NGS Illumina sequencing, e.g. on the Illumina NextSeq 2000 (Illumina, Inc,, San Diego, CA). This approach wi ll result in sequenced fragments that share the sequence of either the RpoB1-R1327 primer or the Cpn60-R571 primer, followed by sequences that vary in length and nucleotide composition. Sequences derived from the same microorganisms and extended from the same primer will be identical except for the length of the sequenced fragment, which will vary as a function of the distance between the respective primer annealing site and the end of the mcfDNA fragment.

[00291 ] The processing and analysis of the SPA fragment sequences can include the following steps:

1 , Similar to single loci SPA fragment sequencing, the reads are filtered based on read quality. Error correction can be done using software such as DADA2 (Callahan et al, 2016), which makes use of a parametric error model. The remaining error-corrected reads of different lengths can be deduplicated while recording the number of duplicates by sequence for calculating community composition.

2. Multi loci SPA fragment sequencing can include a step to deconvolute the reads on the phylogenetic gene level. Unique SPA fragments are aligned on the sequences of the RpoB1-R1327 primer or the Cpn60-R571 primer and sorted in gene specific “buckets”. This is schematically shown in Step 1 of Figure 3B. Subsequently, the sequences of each bucket are sorted into bins of matching sequences representative for the same species. In a next step, the rpoB and cpn60 gene databases are searched for the longest read in each bin of matching sequences for species identification. If a fragment does not match exactly to the database entries, the closest match species is assigned, noting the likelihood of a false match. For each phylogenetic gene, the community composition is calculated based on the percent of reads assi gned to each species, taking into consideration the number of duplicate reads identified in step 1. To reconciliate the outcomes obtained for the SPA fragments obtained from different phylogenetic identifier genes, their results are compared and consolidated into a consensus community description (species and their relative abundances), as is schematically shown in Step 2 of Figure 3B,

REFERENCES

Abadio A. K. R., et al. (2011 ). Comparative genomics allowed the identification of drug targets against human fungal pathogens. BMC Genomics 12:75.

Abdulamir A. S., Hafidh R. R,, Bakar F. A. (201 1). The association of Streptococcus bovis/gallolyticus with colorectal tumors: The nature and the underlying mechanisms of its etiological role. Journal of Experimental & Clinical Cancer Research 30: 1 1.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W. , Lipman, D.J. ( 1990), Basic local alignment search tool. J. Mol, Biol. 215:403-410.

Ananthakrishnan A. N., et al. (2017). Gut Microbiome Function Predicts Response to Antiintegrin Biologic Therapy in Inflammatory Bowel Diseases. Cell Host & Microbe 21 : 603- 610.

Anttila T., et al (2001). Serotypes of Chlamydia trachomatis and risk for development of cervical squamous cell carcinoma. JAMA 285:47-51.

Arahal D. R, (2014). Chapter 6 - Whole-Genome Analyses: Average Nucleotide Identity. Methods in Microbiology 41 : 103-122.

Banerjee S.< et al. (2021). Prognostic correlations with the microbiome of breast cancer subtypes. Cell Death & Disease 12: Article number 831 ,

Baxter N.T., et al. (2016). Microbiota-based model improves the sensitivity of fecal immunochemical test for detecting colonic lesions. Genome Med. 8: Article number 37.

Bhattacharjee S., Lukiw W. J. (2013). Alzheimer's disease and the microbiome. Frontiers in Cellular Neuroscience 7: Article 153.

Boleij A., et al. (2011). Novel clues on the specific association of Streptococcus gallofyticus siihsp ga//olyticus with colorectal cancer. Journal of Infectious Diseases 203: 1101- 1 109.

Boulange C. L., et al, (2016). Impact of the gut microbiota on inflammation, obesity, and metabolic disease. Genome Medicine 8:42

Bullman S,, et al, (2017). Analysis of Fusobacteriwn persistence and antibiotic response in colorectal cancer. Science 358: 1443-1448.

Burnham P., Kim M. S., Agbor-Enoh S., Luikart H., Valantine H. A., Khush K. K., et al. (2016). Single-stranded DNA library preparation uncovers the origin and diversity of ultrashort cell-free DNA in plasma. Sei Rep-Uk. 6.

Callahan, B. J., ei al. (2016). DADA2: High-resolution sample inference from Illumina amplicon data. Nature methods 13: 581-583. Castellarin M., et al. (2012). Fi/sobacfemtm nucleatum infection is prevalent in human colorectal carcinoma. Genome Res. 22: 299-306.

Chandra V. ami McAllister F. (2021 ). Therapeut ic potential of microbial modulation in pancreatic cancer. Gut 70; 1419-1425.

Chen C., et al. (2018). Oral microbiota of periodontal health and disease and their changes after nonsurgical periodontal therapy. The [SME Journal 12:1210-1224.

Cohen R. Shannon B. A., McNeal J. E., Shannon T., Garrett K. L, (2005). Propionibacterium acnes associated with inflammation in radical prostatectomy specimens: A possible link to cancer evolution? J. Urol. 173; 1969- 1974.

Coutinho H. D. M,, Falcao-Silva V. S., Goncalves G. F. (2008), Pulmonary bacterial pathogens in cystic fibrosis patients and antibiotic therapy: a tool for the health workers. Int Arch Med. 1: 24.

De Abreu A. L. P., et ai. (2016). Association of human papillomavirus, Neisseria gonorrhoeae and Chlamydia trachomatis co-infections on the risk of high-grade squamous intraepithelial cervical lesion. Am J Cancer Res 6: 1371-1383,

Decker B., Sholl L. M. (2020). Cell-Free DNA Testing. In: Genomic Medicine. Springer: 41-54.

Depoorter E., et al. (2016), Burkltalderia: an update on taxonomy and biotechnological potential as antibiotic producers. Appl Microbiol Biotechnol 100: 5215-5229.

Dominy S. S,, et al. (2019). Porphyromonas gingivatis in Alzheimer's disease brains; Evidence for disease causation and treatment with small-molecule inhibitors. Sci. Adv. 5: eaau3333

Fang X., Monk J. M., MihN., Du B., Sastry A. V., Kavvas E., Seif Y., Smarr L., and Palsson B. O. (2018). Escherichia coli B2 strains prevalent in inflammatory bowel disease patients have distinct metabolic capabilities that enable colonization of intestinal mucosa. BMC Systems Biology 12: 66.

Fernandez-Carballo B. L., Broger T., Wyss R,. Banaei N., Dcnkinger C, M. (2019). Toward the development of a circulating free DNA-based in vitro diagnostic test for infectious diseases: a review of evidence for tuberculosis, .1 Clin Microbiol. 57: eOl234-18.

Fernandez-Garcia, D„ Hills, A., Page, K. et al. (2019). Plasma cell-free DNA (cfDNA) as a predictive and prognostic marker in patients with metastatic breast cancer. Breast Cancer Res 21: 149. https://doi.org/TO.1 186/S13058-019- 1235-8 Geller L. T. _? et al. (2017). Potential role of intratumor bacteria in mediating tumor resistance to the chemotherapeutic drug gemcitabine. Science 357: 1 156-1 160,

Ghebremedhin B,. Layer F., Konig W., Konig B. (2008). Genetic classification and distinguishing of Staphylococcus species based on different partial gap, 16S rRNA, hsp60, rpoB, sodA, and tuf gene sequences. J Clin Microbiol. 46: 1019 -1025.

Gopalakrishnan V., Weiner B., Ford C. B., Sellman B. R., Hammond S. A., Freeman D. J., Dennis P., Soria J~C., Wortman J. R., Henn M. R. (2020). Intervention strategies for microbial therapeutics in cancer immunotherapy, Immuno-Oncology Technology 6: 9-17.

Gorelick J. I.. Senterfit L. B., Vaughan E. D. Jr. (1988), Quantitative bacterial tissue cultures from 209 prostatectomy specimens: Findings and implications. J. Urol. 139: 57 -60.

Gupta R. S., Lo B, Son J. (2018) . Phylogenomics and Comparative Genomic Studies Robustly Support Division of the Genus Mycobacterium into an Emended Genus Mycobacterium and Four Novel Genera. Frontiers in Microbiology' 9: article 67.

Gupta S., et al. (2019). Amplicon sequencing provides more accurate microbiome information in healthy children compared to culturing. Communications Biology 2: 29 ! .

Gyles C., Boerlin P. (2014). Horizontally transferred genetic elements and their role in pathogenesis of bacterial disease. Veterinary Pathology 51 : 328-340.

Han D., Li R., Shi J., Tan P., Zhang R., Li J. (2020). Liquid biopsy for infectious diseases: a focus on microbial cell-free DNA sequencing. Theranostics 10: 5501-5513.

Head S. R., et al. (2014). Library construction for next-generation sequencing: overviews and challenges. BioTechniques 56, 61-64, 66, 68, passim.

Jahr S., Hentze H., Englisch S., Hardt D.„ Fackelmayer F. O., Hesch R. D., et al. (2001). DNA fragments in the blood plasma of cancer patients: quantitations and evidence for their origin from apoptotic and necrotic cells. Cancer Res. 61: 1659- 1665.

Jin C., et al. (2019). Commensal microbiota promote lung cancer development via y8T cells. Cell 176: 998-1013.

Kanagasingam S., Clnikkapalli S. S, Richard Welbury and Sim K, Singhrao. (2020). Porphyromonas gingivalis is a Strong Risk Factor for Alzheimer's Disease. Journal of Alzheimer's Disease Reports 4: 501—51 I .

Kordahi Ml C., et al. (2021). Genomic and functional characterization of a mucosal symbiont involved in eariy-stage colorectal cancer. Cell Host & Microbe 29: 1—10.

Lancefield R. C. (1933). A serological differentiation of human and other groups of hemolytic streptococci. J Exp Med. 57: 571-595. Liang W., Zhao Y., Huang W., Gao Y., Xu W., Tao J., et al. (2019). Non-invasive diagnosis of early-stage lung cancer using high-throughput targeted DNA methylation sequencing of circulating tumor DNA (ctDNA). Theranostic-s 9; 2056-2070.

Links M. G., Dumonceaux T. J., Hemmingsen S. M., Hill J. E. (2012). The chaperonin- 60 universal target is a barcode for bacteria that enables de novo assembly of rnelagenomic sequence data. PLoS One 7: e49755.

Malinowski B.. et al. (2019). The role of Tannerella forsylhia and Porphyromonas gingivalis in pathogenesis of esophageal cancer. Infectious Agents and Cancer 14:3.

Menzel P., et al. (2016). Fast and sensitive taxonomic classification for metagenomics with Kaiju. Nat. Common. 7: 1 1257.

Alni U. N,. Haley C. T., Tyring S. K. (2017). Viral Oncology: Molecular Biology and Pathogenesis. J. Clin. Med. 6 ( 111 ). doi: 10.3390/jcm612011 1 .

Narunsky-Haziza L., et al. (2022). Pan-cancer analyses reveal cancer-type-specific fungal ecologies and bacleriome interactions. Cell 1.85: 3789- 3806.

Nejman D., et al. (2020). The human tumor microbiome is composed of tumor typespecific intracellular bacteria. Science 368: 973-980.

Nomura J., Rieg G., Bluestone G., Tsai T., Lai A., Terashita D., et al. (2019). Rapid detection of invasive Mycobacterium chimaera disease via a novel plasma-based next- generation sequencing test. BMC Infect Dis. 19: 371 ,

Ogier J.-C., Pages S., Galan ML, Barret M., and Gaudriault S. (2019). rpoB, a promising marker for analyzing the diversity of bacterial communities by amplicon sequencing. BMC Microbiology 19: 171.

Parks ID. HL, Imelfbrt M., Skennerton C. T., Hugenboltz P., Tyson G. W. (2015). CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Research 25: 1043-1055.

Parks D. H., et al. (2018), A standardized bacterial taxonomy based on genome phylogeny substantially revises the tree of life. Nature Biotechnology 36: 996-1004,

Parks D. H.. Chuvoehina ML, Reeves P. R., Beatson S. A., Hugenholtz P. (2021 ). Reclassification of Shigella species as later heterotypic synonyms of Escherichia coll in the Genome Taxonomy Database. BioRxiv: https://doi.Org/10.H0.l/202l.09.22.461432

Pasquereau-Kotula E.. Martins M., Aymeric L., Dram si S. (2018). Significance of Streptococcus gallolyticus subsp. gallolyticus association with colorectal cancer. Frontiers in Microbiology 9: article 614. Peters B. A., Hayes R. B., Goparaju C., Reid C., Pass H. L, Ahn J. (2019). The microbiome in lung cancer tissue and recurrence-free survival. Cancer Epidemiol Biomarkers Prey. 28: 731 -740.

Pleguezuelos- Manzano C., et al. (2020). Mutational signature in colorectal cancer caused by genotoxic pfa* E. coli. Nature 580: 269-273.

Poirier S.. et al. (2018). Deciphering intra-species bacterial diversity of meat and seafood spoilage microbiota using gyrB amplicon sequencing. A comparative analysis with I6S rDNA V3-V4 amplicon sequencing. PLoS One. 13: e0204629.

Poore G. D.» et al. (2020). Microbiome analyses of blood and tissues suggest cancer diagnostic approach. Nature 579: 567 -574.

Rassoulian Barrett S. L., Holmes E. A,, Long D. R., Shean R. C., Bautista G. E., Ravi shankar S.» Peddu V., Cookson B. T., Singh P. K., Greninger A. L., Salipante S. ,1. (2020). Cel l free DNA from respiratory' pathogens is detectable in the blood plasma of Cystic Fibrosis patients. Scientific Reports 10, Article number: 6903.

Riggo M. P. and Lennon A. (2007). Development of a novel PCR assay for detection of Prevotella oris in clinical specimens. FEMS Microbiol Lett. 276: 123 128.

Riquelme E., et al. (2019). Tumor microbiome diversity and composition influence pancreatic cancer outcomes. Cell 178: 795 -806.

Savi D., et al. (2019). Impact of clonally-related Burkholderia contaminant strains in two patients attending an Italian cystic fibrosis centre: a case report. BMC Pulmonary Medicine 19; 164.

Shahir N.M., el al. (2020). Crohn s Disease Differentially Affects Region-Specific Composition and Aerotolerance Profiles of Mucosally Adherent Bacteria. Inflammatory Bowel Disease 26: 1843-1855.

Sepich- Poore G. D., Zitvogel L., Straussman R., Hasty L, Wargo J. A., Knight R. (2021). The microbiome and human cancer. Science 371: 1331-1343.

Shahir N. M., et al. (2020). Crohn's disease differentially affects region-specific composition and aerotolerance profiles of mucosally adherent bacteria. Inflammatory Bowel Disease 26: 1843-1855.

Socransky S. S,, Haffajee A. D., Cugini M.A., Smith C., Kent Jr. R. I... ( 1998). Microbial complexes in subgingival plaque. J Clin Periodontal 25: 134-144.

Sun Y., An K., Yang C. (2019). Circulating cell-free DNA. In: Liquid Biopsy. Intech Open, Thomas A.M., et al. (2019). Metagenomic analysis of colorectal cancer datasets identifies cross-cohort microbial diagnostic signatures and a link with choline degradation . Nat. Med. 25: 667-678.

Urbaniak C., et al. (2016). The microbiota of breast tissue and its association with breast cancer. Appl. Environ. Microbiol. 82: 5039 -5048

Venkataramani V., et al. (2019). Glutamatergic synaptic input to glioma cells drives brain tumour progression. Nature 573: 532-538.

Vos M., Quince C., Fiji A. S., de Hollander M., Kowalchuk G. A. (2012). A comparison of rpoB and 16S rRNA as markers in pyrosequencing studies of bacterial diversity, PLoS One 7 :c30600.

Vudatha V., Ranson M., Blair L., Ahmed A. A, (2019), Rapid detection of bacilli Calmette -Guerin-associatcd mycotic aortic aneurysm using novel cell-free DNA assay. J Vase Surg Cases Innov Tech. 5: 143-148.

Wattam A . R., et al. (2014). PATRIC, the bacterial bioinformatics database and analysis resource. Nucleic Acids Res 42, D581-591, doi:10.1093/nar/gktl099.

Wit R. G., Blair L., Frascoli M., Rosen M. J., Nguyen Q.-H., Bercovici S,, et al. (2020). Detection of microbial cell-free DNA in maternal and umbilical cord plasma in patients with chorioamnionitis using next generation sequencing. PLoS ONE 15 (4): e0231239.

Wood D. E., Salzberg S. L. (2014). Kraken: ultrafast metagenomic sequence classification using exact alignments. Genome Biology 15:R46.

Wood D. E, Jennifer Lu J., Langmead B. (2019). Improved metagenomic analysis with Kraken 2. Genome Biology 20, Article number: 257

Wu J., Hu S., Zhang L., Xin J., Sun C., Wang L., et al. (2020). Tumor cireulome in the liquid biopsies for cancer diagnosis and prognosis. Theranostics 10: 4544-4556.

Zeller G., et ai. (2014). Potential of fecal microbiota for early stage detection of colorectal cancer. Moi. Syst, Biol. 10: 766

Zeng Q., et al. (2019). Synaptic proximity enables NMDAR signalling to promote brain metastasis. Nature 573: 526-531.

Zhou Y., Hemmige V„ Dalai S. C., Hong D. K„ Muldrew K., Mohajer M, A. (2019). Utility of whole-genome next-generation sequencing of plasma in identifying opportunistic infections in HIV AIDS. The Open AIDS Journal 13: 7-11.