PATENT COOPERATION TREATY PATENT APPLICATION

METHODS FOR DETECTING INFECTIONS AND

ASSOCIATED IMMUNE RESPONSES

Inventors: Tasha FINGERUN

Roland MARCUS

Richard MARTIN

Brian O’CONNOR

Kendra WALTON

Assignees: Peak Diagnostic Partners, Inc.

National Jewish Health

Entity: Large

Prepared by: Procopio. Cory, Hargreaves & Savitch ELF

525 B Street, Suite 2200

San Diego, California, U.S.A. 92101

METHODS FOR DETECTING INFECTIONS AND ASSOCIATED IMMUNE RESPONSES

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This Application claims priority to and the benefit of U.S. Provisional Patent Application Serial No. 63/340,831, filed May 11, 2022, and entitled “TOOL FOR DETECTING INFECTIONS AND METHODS OF USING THE SAME”, and U.S. Provisional Patent Application Serial No. 63/465,189, filed May 9, 2023, and entitled “METHODS FOR DETECTING INFECTIONS”, the disclosures of which are incorporated herein by reference in their entirety.

HELD

[0002] This disclosure is in the field of next-generation sequencing (NGS), and in particular in toe field of utilizing NGS for detecting infections through the use of particular panels, as described herein.

BACKGROUND

[0003] The diagnosis of pathogen infections has been developed over the past several decades. In the past, doctors and medical professionals relied mainly on physical symptoms and patient histoiy to make a diagnosis. Illis approach, known as clinical diagnosis, was often imprecise and sometimes led to misdiagnosis or delayed diagnosis. With the advent of new diagnostic technologies and methods, the accuracy and speed of pathogen diagnosis have improved. Common methods for diagnosing pathogen infections is through laboratory testing. There are various laboratory methods available, including culture-based methods, serological teste, and molecular diagnostics.

[0004] Current methods for diagnosis of pathogen infections may require weeks to complete, have low sensitivity, and/or are limited in the scope of pathogens that can be detected. In the absence of a positive result, some physicians still prescribe antibiotics due to a tear of undertreating patients. Other physicians, aware of over-use, do not prescribe antibiotics even though an infection may exist Therefore, improved tools and methods for diagnosing infections and determining appropriate treatment plans are needed. SUMMARY

[0005] In an aspect, a system for detecting infection is disclosed. The system includes a pathogen detection component for detecting a pathogen; a host immune response component for detecting a host immune response associated with the pathogen; and an antimicrobial resistance component for detecting an antimicrobial resistance associated with the pathogen.

[0006] In embodiments, the pathogen detection component detects mRNA for species-specific detection of one or more pathogens selected from the group consisting of bacterial pathogens, viral pathogens, and fungal pathogens. In embodiments, the host immune response component detects a human immune gene expression profile. In embodiments, the antimicrobial resistance component detects anti-microbial resistance gene expression.

[0007] In another aspect, a panel for detecting a pathogen and an associated biological response in an ex vivo biological sample is disclosed. The panel includes a pathogen detection component for amplifying at least a portion of a gene associated with the pathogen; and an immune-response component for amplifying at least a portion of a gene associated with an immune response associated with the pathogen.

[0008] In embodiments, the panel further includes an antimicrobial resistance component for amplifying at least a portion of a gene associated with an antimicrobial resistance to the pathogen.

[0009] In embodiments, the pathogen is a bacteria, a virus, or a fungus. In embodiments, the bacteria is Staphylococcus aureus, Pseudomonas aeruginosa, Haemophilus parainjluenzae, Haemophilus influenzae, Mycoplasma pneumoniae, or a mutant thereof. In embodiments, the fungus is Candida albicans, Aspergillus fumigatus, Candida glabrata, or a mutant thereof. In embodiments, the virus is a Hepatitis virus or a Coronavirus, or a mutant thereof.

[0010] In embodiments, when the bacteria is Staphylococcus aureus, the gene is selected from one or more of dksA l , fecl l 1, fecl_12, fecl_5, iscR, mucA, psIM, pTlF-4, spell, rot, femC, and lpI3_l . In embodiments, when the bacteria is Pseudomonas aeruginosa, the gene is selected from one or more of psiM, mucA, iscR, fecl_5, fecl_12, fecl_11 , ptIF_4, and dksA_l . In embodiments, when the bacteria is Haemophilus parainfluenzae, the gene is selected from one or more of 16s I , I6s_2, and 16s_3. In embodiments, when the bacteria is Haemophilus influenzae, the gene is selected from one or more of bgpC, relB, uppB, vapD, and yafQ l. In embodiments, when the bacteria is Mycoplasma pneumoniae, and the gene is selected from one or more of galU, hisS, and rpol). In embodiments, when the fungus is Candida albicans, the gene is selected from one or mote of MEY 03361, MEYJM223, MEY_05088, and MEY_06242. In embodiments, when the fungus is Aspergillus fumigatus, and the gene is selected from one or more of AFUA 1G05430, AFUA_8G009I0, and AFUA_8G02500. In embodiments, when the fungus is Candida glabraia, and the gene is selected from one or more of CAGLOA00187g, CAGLOA00297g, and CAGL()A(X)803g. In embodiments, when the virus is Coronavirus HKUI, the gene is selected from one or more of Replicase IB and Spike. In embodiments, when the virus is Hepatitis G, the gene is 5’UTR.

[0011] In embodiments, the gene associated with an antimicrobial resistance is selected from one or more of aph(3”)=lb, aac(6>lak, blaPDC, erm(G), norA, and su!2.

[0012] In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 1 -26, or 110- 112. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 27-52, 63-75, or 104-109. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 53-56, or 120- 126. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 57-59. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 60-62. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 76-78. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 79-85. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 86-87, or 116- 118. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 88-103. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 113- 115. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to SEQ ID NO: 119. [0013] In embodiments, the immune-response component is associated with a bacterial pathogen, a viral pathogen, or a fungal pathogen. In embodiments, the immune-response component is associated with a bacterial pathogen, and the gene is selected from TRAF3, PIK3CD, RUNX1, 1FNBL 1FNA5, RORC, IFNG-ASI, IFNA4, IFNA2, and RORA. In embodiments, the immune-response component is associated with a bacterial pathogen, and the gene is selected from KAV'i, PRKCD, ICOSLG, IRAK2. TICAM1, CSF3, TRAF3. AGER, PTGER4, and BCU. In embodiments, the immune-response component is associated with a viral pathogen, and the gene is selected from SOCS1. IL13, CSF2, PRKCD, NLRP3, DDX3X, CCR8, TGFB1, CCL17, and TNTRSF4. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 127-136.

BRIEF DESCRIPTION OF THE FIGURES

[0014] FIG. 1 illustrates an algorithm for predicting infection in accordance with embodiments of the disclosure.

[0015] FIG. 2 illustrates a limit of detection assay in accordance with embodiments of the disclosure.

DETAILED DESCRIPTION

[0016] Overview of the Detailed Descript ton

[0017] The detailed description of exemplary embodiments references the accompanying drawing figures, which show exemplary embodiments by way of illustration and their best mode. While these exemplary embodiments are described in sufficient detail to enable those persons skilled in tlie art to practice the invention, it should be understood that other embodiments may be realized and that logical, chemical, and mechanical changes may be made without departing from the spirit and scope of the inventions detailed herein. Thus, the detailed description is presented for purposes of illustration only and not of limitation. For example, unless otherwise noted, the steps recited in any of the method or process descriptions may be executed in any order and are not necessarily limited to the order presented, Furthermore, any reference to singular includes plural embodiments, and any reference to more than one component or step may include a singular embodiment or step. Also, any reference to attached, fixed, connected or the like may include permanent, removable, temporary, partial, full and/or any other possible attachment option. Additionally, any reference to without contact (or similar phrases) may also include reduced contact or minimal contact.

[0018] A sequencing diagnostic panel is a type of molecular diagnostic test that uses next- generation sequencing (NGS) technology to detect and identify pathogens in patient samples. Disclosed herein is a comprehensive, yet targeted, next-generation sequencing (NGS) diagnostic panel for identifying infections.

[0019] Definitions and Interpretation

[0020] If and as used herein, the term “pathogen detection” refers to a component of: (1) six (6) genes, each of which provides speciation information across a wide range of microbes, such that the combined information enables global detection of potential pathogens; and (2) multiple mRNAs for species-specific detection of bacterial, viral, and fungal pathogens.

[0621] If and as used herein, the term “anti-microbial resistance” refers to a component that detects anti-microbial resistance (AMR) gene expression.

[0022] If and as used herein, the term “host immune gene expression” component measures the human immune/inflammatoiy profile. Results from all components of the panel are derived from the same patient sample in the same assay.

[0023] Overview of Aspects & Embodiments

[0024] In an aspect, a system for detecting infection is disclosed. The system includes a pathogen detection component for detecting a pathogen; a host immune response component for detecting a host immune response associated with the pathogen; and an antimicrobial resistance component for detecting an antimicrobial resistance associated with the pathogen.

[0025] In embodiments, the pathogen detection component detects mRNA for species-specific detection of one or more pathogens selected from the group consisting of bacterial pathogens, viral pathogens, and fungal pathogens. In embodiments, the host immune response component detects a human immune gene expression profile. In embodiments, the antimicrobial resistance component detects anti-microbial resistance gene expression.

[0026] In another aspect, a panel for detecting a pathogen and an associated biological response in an ex vivo biological sample is disclosed, The panel includes a pathogen detection component for amplifying at least a portion of a gene associated with the pathogen; and an immune-response component for amplifying at least a portion of a gene associated with an immune response associated with the pathogen.

[0027] In embodiments, the panel further includes an antimicrobial resistance component for amplifying at least a portion of a gene associated with an antimicrobial resistance to the pathogen.

[0028] In embodiments, the pathogen is a bacteria, a virus, or a fungus. In embodiments, the bacteria is Staphylococcus aureus, Pseudomonas aeruginosa, Haemophilus parainfluenzae, Haemophilus influenzae, Mycoplasma pneumoniae, or a mutant thereof. In embodiments, the fungus is Candida albicans, Aspergillus fumigatus, Candida glabrata, or a mutant thereof. In embodiments, the vims is a Hepatitis virus or a Coronavirus, or a mutant thereof.

[0029] In embodiments, when the bacteria is Staphylococcus aureus, the gene is selected from one or more of dksA_l, fecl_ri, fecl_12, fecl_5, iscR, mucA, psiM, pTIF-4, spell, rot,femC, and lpI3_l . In embodiments, when the bacteria is Pseudomonas aeruginosa, the gene is selected from one ot more of psiM, mucA, iscR, fecl_5, fecl_12, fecl_11 , ptIF_4, and dksA_l . In embodiments, when the bacteria is Haemophilus parainfluenzae, the gene is selected from one or more of 16s_l , I6s_2, and 16s_3. In embodiments, when the bacteria is Haemophilus influenzae, the gene is selected from one or more of hgpC, relB, uppB, vapD, and yafQ^l. In embodiments, when the bacteria is Mycoplasma pneumoniae, and the gene is selected from one or more of galU, hisS, and rpoD. In embodiments, when the fungus is Candida albicans, the gene is selected from one or more of MEY 03361, MEY„04223, MEY .05088, and MEYJ)6242. In embodiments, when the fungus is Aspergillus .fumigatus, and the gene is selected from one or more of AFUA__1G0543O, AFUA_8G00910, and AFUA_8G02500. In embodiments, when the fungus is Candida glabrata, and the gene is selected from one or more of CAGLOA00187g, CAGLOA00297g, and CAGLOA00803g. In embodiments, when the virus is Coronavirus HKU1, the gene is selected from one or more of Replicase IB and Spike. In embodiments, when the virus is Hepatitis G, the gene is 5’UTR.

[0030] In embodiments, the gene associated with an antimicrobial resistance is selected from one or more of aph(3”)~lb, aac(6’)-lak, blaPDC, erm(G), norA, and sul2. [0031] In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 1-26, or 1 10-112. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, WA, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 27-52, 63-75, or 104- 109, In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 53-56, or 120-126. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 57-59. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 60-62. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%. 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 76-78. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91 %, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 79-85. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%. 87%, 88%, 89%, 90%, 91%, 92%. 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 86-87, or 116-118. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91 %, 92%, 93%. 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 88-103. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to any one of SEQ ID NOs: 113- 115. In embodiments, the sequence of the gene is at least 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98% or 99% identical to SEQ ID NO: 119.

[0032] In embodiments, the immune-response component is associated with a bacterial pathogen, a viral pathogen, or a fungal pathogen. In embodiments, the immune-response component is associated with a bacterial pathogen, and the gene is selected from TRAF3, P1K3CD, RUNXI, IFNB1, IFNA5, RORC, 1FNG-AS1, 1FNA4, IFNA2, and RORA. In embodiments, the immune-response component is associated with a bacterial pathogen, and the gene is selected from PAVI. PRKCD, ICOSLG. IRAK2. TIC AMI , CSF3, TRAF3. AGER, PTGER4, and BCL6. In embodiments, the immune-response component is associated with a viral pathogen, and the gene is selected from SOCSI. IL13, CSF2. PRKCD, NLRP3, DDX3X, CCR8, TGFB1, CCLI7, and TNFRSF4. In embodiments, the sequence of the gene is at least 80%, or at least 90%, or at least 95% identical to any one of SEQ ID NOs: 127-136.

[0033] Results from this system and the associated panel can be used to effectively treat an individual based on the appropriate detection of the pathogen. Notably, the disclosed system and panel allow for the rapid and correct determination of a pathogen based on pathogen detection, host immune gene expression, and anti-microbial resistance phenomena.

[0034]

[0035] A sequencing diagnostic panel is a type of molecular diagnostic test that uses next- generation sequencing (NGS) technology to detect and identify pathogens in patient samples. Unlike traditional diagnostic teste, which are often limited to detecting one or a few specific pathogens, sequencing diagnostic panels can simultaneously screen for hundreds or thousands of different pathogens in a single test.

[0036] Ihe process of using a sequencing diagnostic panel typically involves collecting a sample from the patient, such as blood, urine, or tissue, and extracting the genetic material from the sample. The extracted genetic material is then processed and sequenced using NGS technology. The resulting data is analyzed using bioinformatics software, which compares the genetic material to a reference database of known pathogens to identify any matches.

[0037] Sequencing diagnostic panels have several advantages over traditional diagnostic teste. For example, they can detect a wide range of pathogens, including bacteria, viruses, fungi, and parasites, in a single test, litis can be especially useful for identifying rare or emerging pathogens that may not be included in traditional diagnostic panels. Additionally, sequencing diagnostic panels can provide a more accurate diagnosis by detecting genetic variations that may be missed by other tests. However, sequencing diagnostic panels can also be more expensive and time- consuming titan traditional diagnostic teste, and may require specialized equipment and expertise io perform and interpret

[0038] Disclosed herein is a comprehensive, yet targeted, next-generation sequencing (NGS) diagnostic panel for identifying infections, such as microbial pulmonary infections that result in poor symptom control of asthma, COPD and other chronic lung diseases. The innovative design of the panel permits screeni ng for tens of thousands of microbial species and other pathogens in a single assay. The panel also detects anti-microbial resistance genes and human host response genes, enabling characterization of pathogenic activity, microbial longitudinal evolution, and the accuracy of pathogen versus commensal microbial identification. I'he increased sensitivity of the panel may be helpful for detection of various infection types and severity, such that the panel may be a valuable tool for clinical studies of antimicrobial compounds. The panel may also allow researchers to evaluate new treatment paradigms for syndromes, such as chronic lung disease, that target the immune system by studying the interplay between host and microbiome (both pathogenic and commensal) response over time, This application may be of particular value given the likely increased use of biologies to treat chronic lung disease, which can increase the risk of opportunistic infections. Used longitudinally, the panel may assess host and pathogen treatment response simultaneously, thus enabling monitoring for robust investigation of complex infections. Integrating host response with pathogen detection and antimicrobial resistance data may provide evidence of pathogenic microbes, and help measure response to treatinent, shifting microbial activity and/or anti- microbial- resistance evolution.

[0039] The panel may identify pathogens in samples with a low burden of nucleic acids compared to the amount of human nucleic acids. This may occur with less severe infection setting, or in certain other human clinical sample types, such as. for example, synovial fluid, in the context of an acute infection. The panel may use samples directly from the patient, such as bronchoalveolar lavage fluid, sputum, synovial fluid, or tissue biopsy. The panel may screen for hundreds, thousands or hundreds of thousands of bacterial, fungal, and viral species in a single assay by sequencing nucleic acids from genes common to pathogens (e.g., microbial 16s, 18s and ITS (internal transcribed spacer)), but that have variable regions that distinguish among different species.

[0040] The NGS panel consists of three major components. The “Pathogen Detection” component is based on I) six (6) genes, each of which provides speciation information across a wide range of microbes, such that the combined information enables global detection of potential pathogens, and 2) multiple species-specific mRN As for species-specific detection of bacterial, viral, and fungal pathogens. The bacterial detection includes gram stain negative (e.g., Haemophilus influenzae) and gram slain positive species (e.g.. Staphylococcus aureus), mollicute species (e.g„ Mycoplasma pneumoniae), other atypical (do not get colored by gram staining) species (e.g., Chlamydophila pneumoniae), mycobacterial species (tuberculous and non- tuberculous). The fungal detection includes yeast (e.g., Candida albicans) and molds (e.g., Aspergillus fumigates). The viral detection includes single stranded RNA (e.g., Coronavirus HK.U1) double stranded RNA viruses (e.g., Rotavirus), single stranded DNA (e.g., Adeno- associated virus) and double stranded DNA viruses (e.g. Cytomegalovirus). The “Anti-Microbial Resistance” component detects anti-microbial resistance (AMR) gene expression for classes of anti-microbial compounds including the classes aminoglycoside, beta-lactam, fluoroquinolone, macrolide and sulfonamide. The “Host Immune Gene Expression” component measures human immune/infiammatory profiles.

[0041] The NGS panel provides a significant advantage over current testing methodology, as outlined in Table 1 herein.

Table L Comparison of characteristics of existing technologies

[0042] The majority of microbiological clinical testing does not use molecular methods, but instead relies on decades-old approaches for identification of pathogens based on cultures. These culture methods are limited to diagnosis of pathogens that can be grown in the laboratory with known conditions. Furthermore, the microbes that survive the sampling and culturing process may not represent the dominant species or strain driving the infection.

[0043] Some testing employs PCR-based methods. Many PCR assays still use samples from a culture, but there are point-of-care molecular diagnostic assays that use samples directly from a patient, including those from Biofire (https://www.biofiredx.com) and Cepheid (https://www.cepheid.com). These rapid point-of-care implementations of PCR have limits of detection (LoDs) that are substantially higher than PCR is capable of achieving when implemented in a reference laboratory. Even if appropriate LoDs for infections could be achieved with PCR, the challenges for any version of PCR for diagnosis are: ( I ) the low number of pathogens that can be targeted in a single assay (and the low number of targets per pathogen that can be included); and (2) the inability to know whether or not the amplified target truly came from the pathogen of interest This second issue of specificity is particularly relevant because, since PCR does not provide sequence information, there is no way to check whether the detected sequence is from the targeted pathogen or other material. The NGS panel disclosed herein uses sequencing data to confirm an appropriate match between the detected nucleic acid and the pathogen.

[0044] Another previously used testing platform utilizes metagenomic assays. Metagenomic assays that sequence all of the DMA in a sample are attractive because they do not require a priori information about which pathogen(s) to test for and they have the potential to provide a survey of all of the organisms in the sample. However, these methods are generally less sensitive than targeted amplification methods because of the overwhelming amounts of human nucleic acid in the samples compared to the minute amounts of microbe nucleic acid. Because of this lack of sensitivity, current implementations are not considered appropriate for first-line clinical testing. In addition, the data analysis required for meta-genomics approaches is both complex and difficult to optimize. The NGS panel disclosed herein targets the microbe and human genes of interest, allowing both a broad coverage of pathogens and capture of the relevant human expression. This results in increased sensitivity for pathogen detection and simultaneous characterization of the immune response.

[0045] Another diagnostic approach uses human immune gene expression to predict whether an individual has a viral or bacterial infection or a combination of both. In this approach, dozens of genes for each algorithm arc used to predict the type of infection. However, there is no actual detection of the pathogen and therefore no ability to definitively pair the immune response with a pathogen. This approach also requires that testing be done in blood samples.

[0046] CRISPR-based technologies are being developed for diagnostics of infectious disease and may be cheaper on a per-assay basis, but they do not have higher sensitivity than current culture or PCR techniques, and are not yet able to be highly multiplexed limiting their comparability to the presently disclosed sensitive and comprehensive panel, which also includes detection of the host immune response.

[0047] As described above, and detailed below, the NGS panel disclosed herein is a significant improvement over currently employed methods. The panel is designed for use on either the Thermo Ion Next Generation Sequencing (NGS) platform (https://www.fhermofisher.com/order/catalog/producl/A27212) or the Illumina Next Generation Sequencing platform (https://www.illumina.com). However, any NGS sequencing platform may be used.

EXAMPLES

[0048] Example 1. Limit of Detection Assay

[0049] ]ample Collection

[0050] Staphylococcus aureus (also referred to herein as .S'. aureus or “staph*’) has long been recognized as a bacteria that causes disease in humans. It is the leading cause of skin and soft tissue infections such as abscesses (boils), furuncles, mid cellulitis. Although most staph infections are not serious, S. aureus can cause serious infections such as bloodstream infections, pneumonia, or bone and joint infections.

[0051] Pseudomonas is a type of bacteria that is commonly found in the environment, such as in soil or water. Of the many different types of Pseudomonas, the one that most often causes infections in humans is Pseudomonas aeruginosa, which can cause infections in the blood, lungs (pneumonia), or other parts of the body after surgery.

[0052] In this Example, simultaneous detections of both bacterial species were carried out

[0053] Collection of Samples

[0054] Staphylococcus aureus and Pseudomonas aeruginosa were grown from glycerol stocks at 37°C; S aureus in tryptic soy broth (TSB) and P. aeruginosa in lysogeny broth (LB). S. aureus and P. aeruginosa were mixed into one tube at approximately even bacterial numbers (1 x 10 ⁷ of each species). Ten-fold serial dilutions were performed taking 67μl from the original mixed tube using 600μl lx Shield in each dilution. 1 :10 serial dilutions were made to obtain 10 ⁷ to 10 ⁰ dilutions. 10 ^x references the estimated number of organisms in the dilution based on ocular density; estimated number of organisms decreases from 10 ⁷ to 10 ⁰ . Once samples were diluted, 6μl of human sputum was added to every dilution so that the human sputum stayed constant as the bacterial strains were diluted out in lx Shield. In a concurrent experiment, the same serial dilutions were made in PBS, removing the lx Shield as a buffer, so that this dilution series could be plated on an appropriate medium for S. aureus and P. aeruginosa growtii. For each of the 8 dilutions, 20μ of the PBS plus bacterial mixture was plated on 3 plates; one with specific growth factors for Pseudomonas^ one for Staphylococcus and one non-specific LB plate. Each pathogen was considered detected at a given dilution point if there were at least 3 colony forming unite (CPU) observed on the plate, a liberal criteria compared to usual clinical testing (see, for e.g.: O’Toole et al. (2016) J. Bacterial.). P. aeruginosa was detected at the 10 ⁷ 10 ⁶ , 10 ^s , and 10 ⁴ dilutions, but not detected at lower dilution points; therefore, the LoD based on culture was 10 ⁴ for P. aeruginosa. S. aureus was detected at the 10 ⁷ , 10 ⁶ , 10 ⁵ , 10 ⁴ , 10 ³ dilutions, but not detected at lower dilution points; therefore, the LoD based on culture was 10 ³ for S. aureus.

(0055] Extraction of Nucleic Acid

[0056] Samples were thawed on ice and vortexed to mix well. 2ml screw top tubes were filled -1/3 full with 1mm Zirconia/Silica beads. After samples were transferred to the 2ml screw top lubes they were placed in a MP Bio FastPrep-24 classic bead beating grinder and lysis system. The samples were bead beat at 6.5m/s for 45s for 2 rounds. The samples were centrifuged at >8, 000g for 5 min to pellet the beads and the supernatant for each sample was transferred to a new 1.7ml tube. The samples were treated with ProK and incubated at room temperature (RT) for 30 mins. Using the Zymo Quick DNA/RNA Miniprep kit, following the DNA and RNA purification protocol, I volume of Lysis buffer was added. The entire sample was transferred to the Spin- A way Filter and centrifuged at > 8,000g for 30s. The Spin-Away Filter was transferred to a new collection tube while the RNA was processed from the flow through. One volume of ethanol was added to the RNA flow through and transferred to a Zymo-Spin I1CR column and centrifuged at full speed for 30s. DNase treatment was performed on all samples, followed by a DNA/RNA Prep Buffer wash and two DNA/RNA washes, and eluted in 50μl of DNase/RNase free water. The DNA was then extracted by adding DNA/RNA Prep Buffer to the Spin-Away Filter, followed with two rounds of washes. The samples were eluted in 50μl DNase/RNase free water.

(0057] Sequencing Library Preparation [0058] Ion Torrent libraries were prepared using a Therrmo Fisher custom assay and the Ion

Torrent Library Plus kit and protocol (ThermoFisher. Accessed 3.2822, https://www.thermofisher.com/documenl-connect/document- connect.html?url™https%3A%2F%2Fassets.thennofisher.com%2Fr FS- Assets%2FLSG%2FmanuaIs%2FMANOO17003 JonAmpliSeqLibaryKitPlus_UG.pdf).

[0059] RNA from mixed pathogen sputum dilution series samples (100ng) was reverse transcribed using SuperScript I V V I LO Master M ix. For each cDN A sample, PCR master mix was added to 10μl of cDNA. Each sample was amplified for 16-20 cycles with a 4-16 minute anneal/extend time according to protocol After digestion of the primer sequences, barcoded adapters were added to each sample, each sample was cleaned using Ampure XP beads, and the beads were eluted with a library amplification master mix. Libraries underwent a brief PCR, were cleaned with a double Ampure XP bead clean up eluted in 50μl Low TE. Libraries were quantified and pooled based on concentration to obtain equal representation of each sample. Each pool was templated on the Ion Chef Instrument and sequenced on the Ion GeneStudioTM S5 Prime System.

[0060] NGS Pathogen Gene Detection Analysis

[0061] To determine whether and which bacterial, fungal and/or viral species are present in a given sample based on the NGS panel, the algorithm depicted in FIG. 1 was applied to the raw sequencing data based on the library for that sample. Specifically, with respect to FIG. 1, the raw sequencing files were filtered to keep reads that were at least 100 bp long. Reads were aligned to the database of expected sequences based on the NGS panel design using Bowtie2 with ‘Very* sensitive setting (see, for e.g.: Langmead el al. “Scaling read aligners to hundreds of threads on general-purpose processors.” Bioninformalics). Reads that aligned with Bowtie2 were considered potential matches to a target gene sequence in the NGS panel design (target), and those read sequences were then compared to the nucleotide nt database (Nucleotide [Internet], Bethesda (MD): National Library of Medicine (US) NCBI) to determine the single or multiple species consistent with the read’s sequence, noting the % identity of each mapping event. Detection of a bacterial or fungal species in a sample required at least 3 gene targets of that species present in that sample. Detection of a viral species in a sample required at least one gene target of that species be detected in the sample given the much smaller genome sizes. For all pathogens, for a target to be detected for a given species, the reads for that target were required to have >97% identity to the intended species and no greater than 97% identity to any sequence in the database for any other species. The number of reads supporting each target was required to be at least 20 CPM to be considered detected above background. These criteria enhance specificity by requiring a close match of the detected sequence to a specific species with no alternative species being consistent with the target as is described next

(0062] Tables 2 and 3 display, for each of S. aureus and P. aeruginosa at each dilution point, 4 and 8 of the gene targets that were detected for that pathogen, respectively. 5. aureus was detected by the NGS panel at all dilution points but 10 ¹ based on at least 3 gene targets being detected at each dilution point (Table 2). Since S. aureus was not detected at 10 ¹ , the NGS panel LoD for S. aureus was 10 ² even though the panel did detect it at 10 ⁰ . P. aeruginosa was detected by the NGS panel in dilutions 10 ⁷ through 10 ² , but not at dilutions 10’ and 10 ⁰ based on the requirement that at least 3 gene targets are detected (Table 3); dilutions 10’ and 10 ⁰ had no gene targets detected. Hence, the NGS panel LoD for P. aeruginosa was also 10 ² . No other pathogens were detected in any of the dilution series samples such that specificity was 100% across the entire dilution series for both pathogens. The data in Tables 2 and 3 also demonstrate the utility of having more than one gene target available for detection in the panel to increase sensitivity while maintaining specificity; different dilution points can have different genes detected vs. not detected. For example, as shown in Table 2, dilutions 10 ⁶ -10 ² all had the same 4 genes detected (femC, lpI3„l, rot, speH), while the 10 ⁷ dilution had a subset of 3 of those 4 genes (femC, rot, speH) and 10 ⁰ had a different subset of 3 genes (lpI3_l, rot, speH). Similarly, as shown in Table 3 for P. aeruginosa , in dilutions 10 ⁴ and 10 ⁵ , P. aeruginosa was detected based on two of the same genes (mucA and pslM) in each, but 10 ⁴ had the fecl_12 gene as the third detected target whereas 10 ⁵ had the iscR gene as the third detected target. Dilutions 10 ³ (dksA_l, fecl_12, mucA) and 10 ² (fecl_11, fecl_5, ptlF_4) had other combinations of three genes detected. These three dilution points show that by having more than three potential targets in the panel for each pathogen, the sensitivity of the panel is robust to biological differences (e,g., different genes included in different strains of a pathogen genome) or sampling-derived differences (eg., when taking a sample, certain nucleic acids may not be included in the sample) between samples. At the same time, by requiring three different targets, each with uniquely high sequence similarity to the pathogen, the panel demonstrates specificity (low false positive rate) for pathogen calls.

Table 2: Stap/iylococciis aureus Genes Detected at each Dilution Point

Table 3. Pseudomonas aeruginosa Genes Detected at each Dilution Point

[0063] We used DNA from an ATCC mock bacterial community (ATCC MSA- 1002) to determine the analysis pipeline thresholds for determination of whether or not a given species was detected in a sample based on the NGS reads. The mock community included 8 pathogens targeted by our panel (Staphylococcus aureus, Pseudomonas aeruginosa, Escherichia coli, Culibaclerium acnes, Acinetobacter baumannii, Streptococcus agalactiae, Enterococcus faecalis, Staphylococcus epidermis). The samples were prepared as in Example I for library preparation, sequencing, and pathogen detection analysis, where the pathogen detection analysis was modified as described below to investigate a range for 1) the number of reads (in counts per million (CPMj) per target observed that are necessary to determine that the target is detected and 2) the number of targets per species that are necessary to determine that the species is detected. We considered combinations of between 5 CPM and 50 CPM as a threshold for target detection and between 1 and 10 targets detected per species as a threshold for species detection. We compared the sensitivity (probability of cal ling a species as detected when present in the sample) and number of false positive species detected at each combination of CPM and number of targets required. The optimal trade-off between sensitivity and number of false positive detections was identified as either of the following: A) 20 CPM with 2 or 3 targets per species or B) 15 CPM with 3 targets per species required. The sensitivity of the combination of 20 CPM and 3 targets required for bacterial or fungal species detection was 87.5% (7 of 8 detected) with zero false positive findings. No other combination had higher sensitivity without introducing false positive results.

[0064] For each of the genes and at each dilution point, Tables 4 and 5 show the sequence of the gene target from the sequencing data that was detected, the percentage of sequence similarity of the sequence to the reference pathogen sequence in the nt database (% identity), and the CPM for the detected gene target for S aureus and P. aeruginosa, respectively. For each of the genes, the sequence similarity to the reference pathogen sequence is 100% with the exception of one gene that has 99.5% sequence identity. Importantly, even for that gene with 99.5% sequence identity, no other species had a reference sequence with higher than 90% sequence identity to the sequence we identified. These sequence similarities are key for the specificity of the panel performance; target sequences must be identified with very high sequence similarity to a reference sequence that is unique to the pathogen in order to determine that a pathogen is detected, which increases specificity and allows exclusion of pathogens with sequences less similar to the sequence identified in a sample.

Table 4. Sequences Detected for each Gene Target for Staphylococcus aureus in Dilution Series

Table 5. Sequences Detected for each Gene Target for Pseudomonas aeruginosa

[0065] Thereafter , Quantitative Polymerase Chain Reaction Quantitative PCR (qPCR) was performed on the mixture of human sputum, P. aeruginosa and S'. aureus at all 8 dilution points. A No Template Control (NTC) was also included. cDNA synthesis was set up and performed for each sample based on the specific requirements of each qPCR assays needs. [0066] The Pseudomonas-specific RT reactions were set up using 8μl of each dilution point and 2μl of SuperScript IV Vilo Master Mix. For the No RT reaction, Ing (8μl) of sample was used as a template with 2μl SuperScript IV Vilo No RT Ctrl. The samples were run on the thermocycler at 25°C for 10 min, 50°C for 10 min, 85°C for 5 min.

[0067] The S aureus-specific RT reactions were set up using 9.6μl of each dilution point and 2.4μl of Superscript IV VILO Master Mix based on the manufacturer protocol. For the No RT reaction, Ing (9.6μl) of sample was used a template with 2.4μl SuperScript IV Vilo No RT Ctrl. The samples were run on the thennocycler at 25°C for 10 min, 50°C for 10 min, 85°C for 5 min and held at 10°C for less than an hour.

[0068] The cDNA samples were taken to qPCR. ForP. aeruginosa, TaqMan Fast Advanced Master Mix containing ROX (Catalog #4444556) and a Pseudomonas TaqMan assay (20x) were used. Briefly, for evety dilution point, 10μl o f th eaqMan Fast Advanced Master Mix (2x) was mixed with lul of the Pseudomonas TaqMan Assay (20x), 7μl of nuclease free water and 2μl of cDNA. Each dilution point was run in triplicate. The plate was run on the QuantStudio 7 following Pub. No. 4444605, Rev. D.

[0069] Briefly, PCR was set up as follows: Step (UNG incubation); temperature: 50°C; time: 2 mins. Step (polymerase activation); temperature 95°C; time: 2 mins. Step (denature); temperature 95°C; time: 1 second. Step (anneal/extend); temperature 60°C; time: 20 seconds. The latter two steps (i.e., denature and anneal/extend) were repeated for 40 cycles.

[0070] For S’, aureus the TaqMan Staphylococcus aureus Detection Kit (Catalog # 4368606) was used. For every dilution point, 3μl of 10X S aureus Target Assay Mix was mixed with I 5ul 2X Environmental Master Mix and 12μl of cDNA. Each dilution point was run in triplicate.

[0071] The plate was run on the QuantStudio 7 following Pub. No. 4370454, Rev. C. (protocol accessible at: https://assets.thermofisher.cn/ ^, rFS-Assets/LSG/manuals/4370454.pdf, the contents of which are incorporated herein by reference).

[0072] Briefly, PCR was set up as follows: Step 1; temperature 95°C; time: 10 mins. Step 2; temperature 95°C; time: 15 seconds. Step 3; temperature 60°C; time: I minute. The latter two steps (i.e., Steps 2 and 3) were repeated for 40 cycles.

[0073] Detection for qPCR was based on a mean cycle threshold (Ct)<35. [0074] With respect to FIG. 2, the two species in a mixture with human interference (complex sample): S. aureus (left panel) and P. aeruginosa (right panel) in the culture diluted in human sputum. For each panel, the horizontal axis represents the dilution point, with the highest pathogen count on the left and the lowest on the right. For each panel, the left side vertical axis is the cycle threshold (CF) mean of the three replicates for foat dilution point from the qPCR assay; the lower the CF value, the more nucleic acid is present. The solid horizontal blue line (at y-axis value: ~35) denotes a liberal threshold for determining detection of the pathogen based on the CT value (see, for e.g.: Front Microbiol. 2022; 13: 820431 ); when the CT value falls below foat line, the qPCR assay would result in non-detection of the pathogen. For each panel, the right side vertical axis is the number of targets detected by the NGS panel. The solid horizontal orange line (at y-axis value -3) denotes 3 targets detected, the requirement for detection by the NGS assay; when the number of targets detected falls below that line, the NGS panel would result in non-detection of the pathogen. The LoDs for each method (culture, qPCR, NGS panel) as described above and displayed in FIG. 2 show that the NGS panel had at least an order of magnitude lower LoD (and therefore higher sensitivity) than either culture or qPCR while maintaining perfect specificity (no false positive findings for other pathogens).

[0075] Overall, the panel demonstrated a much lower LoD than culture or qPCR while maintaining specificity.

[0076] Example 2. Detection of Fungal Nucleic Acid [0077] Nucleic acid for three fungal species was purchased from the American Type Culture Collection (ATCC): Candida albicans (ATCC- 1023 ID-5), Aspergillus fumigatus (ATCC-MYA- 4609D-2), and Candida glabrata (ATCC-2001D-5). The samples were resuspended from lyophilized pellets to a concentration of 20ng/μl according to their Certificate of Analysis from ATCC. The samples were diluted to ~4ng/μl and pooled at equal volumes of 50μl.

[0078] Sequencing library preparation was conducted as described in Example 1 starting from 20ng DNA rather than RNA. Similarly, Pathogen Gene Detection Analysis was conducted as described in Example I .

[0079] All three fungal species, and no other microbes, were detected in the fungal species pool. Table 6 shows genes for each of Candida albicans (4 genes), Candida glabrata (3 genes), and Aspergillus jumigatus (3 genes), respectively, that were detected by application of the pathogen detection algorithm to the sequencing panel data. Table 7 {Candida albicans), Table 8 (Aspergilius fiunigatus), and Table 9 {Candida glabrata) show the detected sequence from the sequencing data and the CPM for the detected sequence for each gene for each species, respectively. Collectively, these tables demonstrate that multiple targets with extremely high sequence similarity to the species called by the pathogen detection algorithm were detected for each of the three species that were present in the single sample (fungal pool) that contained all three of the fungal species. We did not detect any other species in the pool such that the approach yielded 100% sensitivity and 100% specificity for the three fungal species.

Table 6: Genes Detected for each Fungal Species in the Fungal Pool

[0080] Table 7 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity), and the counts per million (CPM) of that sequence in the NGS sequencing data by the pathogen detection algorithm for each Candida albicans gene in Table 6.

Table 7, Candida albicans Sequences Detected for each Gene Target in the Fungai Pool

[0081] Table 8 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data by the pathogen detection algorithm for each Aspergillus jumtgatus gene in Table 6.

Table 8. Aspergillus fumigatus Gene Target Sequences Detected in the Fungal Pool for each Gene

[0082] Table 9 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data by the pathogen detection algorithm for each Candida glabrata gene in Table 6.

Table 9. Candida glabrata Gene Target Sequences Detected in the Fungal Pool for each Gene

[0083] Example 3. Comparison to Clinical Microbiology Results in Asthma Clinical Samples

[0084] Samples from asthma patients in a clinical research study were used to demonstrate the performance of the panel To be enrolled and have samples taken, the patients had to be free of any symptoms (e.g„ fever) indicative of a respiratory infection. At yearly visits, each patient underwent a bronchoscopy for collection of bronchoalveolar lavage (BAL) fluid in addition to induced sputum collection. The BAL samples were sent to a clinical laboratory for culture and PCR testing. At quarterly clinic visits intervening between the yearly visit that included bronchoscopy, an induced sputum was collected. A total of 97 sputum samples tor 17 patients were tested; patients had between 1 and 9 samples each. Of the 17 patients with at least one induced sputum sample available for study, 6 had both bronchoalveolar lavage (BAL) fluid and sputum samples collected at the same clinical visit.

[0085] Nucleic acid extraction was performed by the method described for the limit of detection assay in Example 1 , above.

[0086] To compare clinical microbiology pathogen testing results to the detection results of the NGS panel, the BAL clinical testing results were compared to the panel pathogen detection results in sputum. To obtain the pathogen detection via the panel, the pathogen detection analysis was applied to the NGS reads from the sputum samples. The concordance or discordance between the BAL clinical pathogen detection and the panel detection was recorded.

(0087] The comparison of the clinical lab pathogen detection versus the panel detection is shown in Table 10. For each of the 5 positive BAL clinical lab results, the panel detected the pathogen(s) indicated by the clinical lab method (culture or PCR), with the exception of one patient, who had low neutrophils and only oral species detected by the panel (Patient E). Given the sensitivity of the panel to detect S. aureus, demonstrated above, and low BAL % neutrophils, it is concluded that the clinical lab results for Patient E may have been incorrect In a patient who was negative for pathogen detection in the clinical lab results (Patient C), the panel detected H. parainfluenzae in sputum.

Table 10. Comparison of Clinical Pathogen Defection in BAL by culture or PCR vs. NGS Panel Pathogen Detection in Sputum [0088] Table II shows 8 genes for Pseudomonas aeruginosa and whether or not they were detected in each of the patient samples listed in Table 10. Patients A, 13 and D all had Pseudomonas aeruginosa detected, but with different combinations of gene targets contributing to that detection; Patient A had three of the eight gene targets detected (fecl_5, mucA dksA_l), Patient B had four of the eight gene targets detected (fecl_5, mucA, psIM, dksA_I ) and Patient D had six of the eight of the gene targets detected (fecl_1 , fecl_5, mucA, psIM, dksA_1 , iscR). None of the other patients in Table 10 had any of those eight gene targets detected fot Pseudomonas aeruginosa.

Table 11. Genes Detected or Not from the NGS Panel for hendomonas aeruginosa in Patients from Table 10

[0089] As generated by the pathogen detection algorithm, Table 12 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for each Pseudomonas aeruginosa gene target detected in Table II. Each of the sequences had 100% sequence identity with the Pseudomonas aeruginosa target sequence in the database.

Table 12. Pseudomonas aeruginosa Gene Target Sequences Detected in Patients from Table 10.

[0090] Table 13 shows 3 gene targets for Haemophilus parainfluenzae and whether or not they were detected in each of the patient samples listed in Table 10. Only Patient C had Haemophilus parainfluenzae detected via detection of all three of the 16s gene targets; none of the other samples had any of those three gene targets for Haemophilus parainfluenzae detected.

Table 13, Targets Detected or Not from the NGS Panel for Haemophilus parainfluenyu in Patients from Table 10

[0091 ] As generated by the pathogen detection algorithm, Table 14 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for each Haemophilus parainfluenzae gene target detected in Table 13. Two of the sequences had 100% sequence identity with the Haemophilus parainfluenzae target sequence in the database and one had 99.3% sequence identity with the Haemophilus parainfluenzae target sequence in the database.

Table 14. Haemophilus perainflueinae Gene Target Sequences Detected in Patients from Table 10.

[0092| Table 15 shows 5 genes for Haemophilus influenzae and whether or not they were detected in each of the patient samples listed in Table 10. Patients B and D both had Haemophilus influenzae detected, but with different combinations of gene targets contributing to that detection; Patient B had three of the five gene targets detected (hgpC, relB, uppB) and Patient D had an additional gene target detected for a total of four of the five gene targets detected (hgpC, relB, uppB, vapD). None of the other patients in Table 10 had any of those five gene targets detected for Haemophilus influenzae.

Table 15. Genes Detected or Not from the NGS Panel for Haemophilus influenzae in Patients from Table 10 [0093] As generated by the pathogen detection algorithm, Table 16 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per mill ion (CPM) of that sequence in the NGS sequencing data for each Haemophilus influenzae gene target detected in Table 15. Each of the sequences had 100% sequence identity with the Haemophilus influenzae target sequence in the database.

Table 16. Haemophilus influenzae Gene Target Sequences Detected in Patients from Table 10.

(0094] Table 17 shows 2 gene targets for Coronavirus HKU1 and whether or not they were detected in each of the patient samples listed in Table 10. Only Patient C had Coronavirus HKU 1 detected via detection of both of the gene targets; none of the other samples had any of those 2 gene targets for Coronavirus HKU1 detected.

Table 17. Genes Detected or Not from the NGS Panel for Coronavirus HKU1 in Patients from Table 10 ♦With reference to Table 17, “DET" means detected: and “ND” means not detected.

[0095] As generated by the pathogen detection algorithm. Table 18 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for each Coronavirus HKUl gene target detected in Table 17. One of the sequences had 100% sequence identity with the Coronavirus HKU 1 target sequence in the database and the other had 99.5% sequence identity with the Coronavirus HKU 1 target sequence in the database.

Table 18. Coronavirus HKUl Gene Target Sequences Detected in Patients from Table 10.

[0096] Collectively, these data show high concordance between Clinical Pathogen Detection and the NGS panel with increased sensitivity and specificity of the NGS panel compared to Clinical Pathogen Detection. All but two of the sequences detected had 100% sequence identity to a reference sequence from the relevant pathogen; the remaining 2 had at least 99.3% sequence identity to a reference sequence.

[0097] Example 4. Pathogen Detection, Immune Response and Antimicrobial Resistance in Asthma Patients

[0098] To demonstrate the improved diagnostic function of the NGS panel, as well as the use of the panel to inform therapeutic interventions, the NGS panel was used to detect multiple pathogens (e.g.,, bacteria, viruses, and fungi), as well as antimicrobial resistance genes, and human immune response genes.

[0099] Anti-Microbial Resistance Gene Detection Data Analysis

[0100] To determine whether or not anti-microbial resistance (AMR) genes are expressed in a sample, a very similar algorithm to the pathogen detection analysis was applied. Specifically, starting with the sequencing file that retains only reads that are at least 100bp long, reads were aligned to the database of expected AMR sequences based on the NGS panel design using Bowtie2 with “very-sensitive” setting, as described and referenced previously herein. Reads that aligned with Bowtie2 were considered potential matches to a target gene sequence in the NGS panel design (target), and those read sequences were then compared to reads that aligned to an AMR gene in the NGS panel design. The total number of reads mapping to a given AMR gene was recorded. An antibiotic class was considered detected in a given sample if at least one gene target from that class was detected (i.e., had at least 20 CPM reads) in that sample, The choice of a subset of the AMR genes in our panel was based on presence of that gene in validated databases such as the National Database for Antibiotic Resistant Organisms (NDARO) with appropriate supporting documentation. We found that six of these genes were ubiquitously expressed in all 97 of the sputum samples, suggesting that they are not appropriate differentiators of AMR activity based on expression levels in sputum. The six genes ubiquitously expressed were cfxA, erm(B), erm(F), erm(X), meRA) and msr(D). These genes were removed from the AMR detection algorithm despite external evidence for their relevance given these data.

[0101] Host Immune Gene Detection Data Analysis

[0102] To determine the read counts for each human immune response gene in the panel, the human immune response CPM counting algorithm started with the filtered sequencing files described herein. A read was retained if it aligned to a human immune target gene sequence (target) that is consistent with the NGS panel design using the same aligner and parameters as used for the Pathogen Detection algorithm and the AMR algorithm, as described herein. The number of reads going to each target was recorded and then standardized by the total number of reads mapping to the human immune portion of the panel to generate the counts per million (CPM) for that gene; this read count table was the input to the immune profile algorithms described below.

[0103) The pathogen detection and AMR detection analysis described above was also applied to the NGS reads from each of the sputum samples from 97 asthma patients. The species detected and antibiotic gene classes detected were recorded for each species.

[0104] Based on the reference set of 97 sputum samples from the asthma patients, patterns of gene expression that may distinguish samples with and without certain pathogens (or pathogen types) were identified. First, for each gene, the gene expression value for each sample (counts per million) was standardized by subtracting the mean and dividing by the standard deviation of the expression counts across the 97 sputum samples. The median of the standardized expression from the genes contributing to a given pattern was calculated to obtain a score for each sample. To determine which samples were predicted by the algorithm to have a given pathogen or pathogen type, a cut-off value of 1.0 was applied to the scores; samples with a score >1.0 were considered to have presence of an immune response to that pathogen or pathogen type.

[0105] To distinguish samples with an immune response to H. influenzae from those without, 10 genes from the immune panel were used (TRAF3, PIK3CD, RL'NXZ, [FNBl. IFNA5, RORC. IFNG-ASi, IFNA4. 1FNA2. RORAy, the median of the standardized expression values tor the 10 genes are used to calculate a score. Table 18A shows each of the genes, their standardized score, and tiie H. influenzae score (median of those values). Similarly, 10 genes each were used for a gram negative bacterial immune response score (K/1F7, PRKCD, ICOSLG, IRAK2, TICAMI, CSF3, TRAF3, AGER, PTGER4, BCL6). Table 18B shows each of the genes, their standardized score, and the Gram negative bacterial immune response score (median of those values). Similarly, 10 genes were used to define a viral response score ^SOCSl, 1L13, CSF2, PRKCD, NLRP3, DDX3X. CCR8. TGFB1, CCL17, TNFRSF4). Table 18C shows each of the genes, their standardized score, and the Viral response score (median of those values). Note that for each of these scores, patients can have similar scores based on different combinations individual gene standardized values.

Table 18 A. Individual Gene Values Contributing to H, influenzae Immune Score Table 18B. Individual Gene Values Contributing to Gram Negative Bacterial Immune Response Score

Table 18C. Individual Gene Values Contributing to Viral Response Immune Score

[0106] Table 19 shows a subset of results from two of the three components of the panel (Pathogen Detection, Immune Response) for 11 sputum samples. Shaded cells indicate that the panel found evidence for the pathogen or that type of immune response. Table 19. Paired Pathogen Detection and Immune Response Score for 11 Patients

[0107] Table 20 shows 5 gene targets for Huemophilus influenzae and whether or not they were detected in each of the patient samples listed in Table 19, Four patients had Haemophilus influenzae detected based on at least three of the five gene targets being detected. Patient 1 had three genes detected (hgpC, relB, vapD). Patients 2 and 4 each had the same four of the five genes detected (hgpC, relB, uppB, vapD), while Patient 3 had ail five of the five genes detected (hgpC, relB, uppB, vapD, yafQ I). None of the other samples had any of those five gene targets for Haemophilus influenzae detected.

Table 20. Haemophilus influetnae Gene# Detected or Net in Patients in Table 19

[0108] As generated by the pathogen detection algorithm, Table 21 shows the target sequence detected, tile percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for each Haemophilus influenzae gene target detected. Each of the sequences had 100% sequence identity with the Haemophilus influenzae target sequence in the database.

Table 21. Haemophilus Influenzae Gene Target Sequences Detected for Patients in Table 19

[0109] Table 22 shows 8 gene targets for Pseudomonas aeruginosa and whether or not they were detected in each of the patient samples listed in Table 19. One patient had Pseudomonas aeruginosa detected based on at least three of the 8 gene targets being detected. Patient 4 had six of the eight genes detected (fecl_5, fecl_12, mucA, psIM, dksA I, iscR). None of the other samples had any of those eight gene targets for Pseudomonas aeruginosa detected.

Table 22. Pseudomonas aeruginosa Genes Detected or Not in Patients in Table 19

[0110] As generated by the pathogen detection algorithm, Table 23 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for each Pseudomonas aeruginosa gene target detected. Each of the sequences had 100% sequence identity with the Pseudomonas aeruginosa target sequence in the database.

Table 23. Pseudomonas aeruginosa Gene Target Sequences Detected for Patients in Table 19

[0111] Table 24 shows 4 gene targets for Staphylococcus aureus and whether or not they were detected in each of the patient samples listed in Table 19. One patient had Staphylococcus aureus detected based on at least three of the four gene targets being detected. Patient 5 had three of the four genes detected (femC, rot, speH). None of the other samples had any of those four gene targets for Staphylococcus aureus detected.

Table 24. StaphylococcMS aureus Genes Detected or Not in Patients in Table 19

[0112] As generated by the pathogen detection algorithm. Table 25 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for each Staphylococcus aureus gene target detected. Each of the sequences had 100% sequence identity with the Staphylococcus aureus target sequence in the database.

Table 2S. Staphylococcus aureus Gene Target Sequences Detected for Patients in Table 19

[0113] Table 26 shows 3 gene targets for Mycoplasma pneumoniae and whether or not they were detected in each of the patient samples listed in Table 19. One patient. Patient 6, had Mycoplasma pneumoniae detected based on all three of the gene targets being detected (galU, hisS, rpoD). None of the other samples had any of those three gene targets for Mycoplasma pneumoniae detected.

Table 26. Mycoplasma pneumoniae Genes Detected or Not in Patients in Table 19

[0114] As generated by the pathogen detection algorithm, Table 27 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for each Mycoplasma pneumoniae gene target detected. Each of the sequences had 100% sequence identity with the Mycoplasma pneumoniae target sequence in the database.

Table 11. Mycoplasma pneumoniae Gene Target Sequences Detected for Patients in Table 19

[0115] Table 28 shows 2 gene targets for Coronavirus HKU1 and whether or not they were detected in each of the patient samples listed in Table 19. One patient had Coronavirus HK.U1 detected based cm at least one of the two gene targets being detected. Patient 8 had both genes detected (repticase IB, spike). None of the other samples had either of those 2 gene targets for Coronavirus HKU1 detected.

Table 28. Coronavirus HKU1 Genes Detected or Not in Patients in Table 19

[0116] As generated by the pathogen detection algorithm, Table 29 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data tor each Coronavirus HKUl gene target detected. Each of the sequences had at least 99.5% sequence identity with the Coronavirus HKUl target sequence in the database.

Table 29. Coronavirus HKUl Gene Target Sequences Detected for Patients in Table 19

[0117] Table 30 shows the target for Hepatitis G and whether or not it was detected in each of the patient samples listed in Table 19. One patient, Patient 9, had Hepatitis G detected based on the 5’ UTR being detected. None of the other samples had that target for Hepatitis G detected.

Table 30. Hepatitis G 5’UTR Detected or Not in Patients in Table 19

[0118] As generated by the pathogen detection algorithm, Table 31 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for the Hepatitis G target detected. That sequence had 100% sequence identity with the Hepatitis G target sequence in the database. Table 31. Hepatitis G S’UTR Gene Target Sequences Detected for Patients in Table 19

[0119] Table 32 shows 4 gene targets for Candida albicans and whether or not they were detected in each of the patient samples listed in Table 19. Two patients had Candida albicans detected based on at least three of the four gene targets being detected. Patient 5 had three of the four genes detected (MEYJB361, MEY_04223, MEY„06242). Patient 7 had four of the four genes detected (MEY_03361, MEYJ14223, MEY„05088, MEY„06242). None of the other samples had any of those four gene targets for Candida albicans detected.

Table 32. Candida albicans Genes Detected or Not in Patients in Table 19

[0120] As generated by the pathogen detection algorithm, Table 33 shows the target sequence detected, the percent sequence identity of that sequence to the reference pathogen sequence in the nt database (% identity) and the counts per million (CPM) of that sequence in the NGS sequencing data for the Candida albicans target detected. Each of the sequences had 100% sequence identity with the Candida albicans target sequence in the database.

Table 33. Candida albicans Gene Target Sequences Detected for Patients in Table 19

[0121] For the 7 patients with a bacterial species detected, Table 34 shows the results from the antimicrobial resistance component of the panel. Shaded cells indicate that genes associated with resistance to a particular antibiotic class were detected in that sample. Five of the 7 patients had evidence of AMR for at least one class of antimicrobial treatment. Table 35 shows a gene used to detect AMR to a given class for each patient in Table 34 and whether or not it was detected in each patient. Table 36 shows the specific sequence for each of the gene targets detected for each of the patients with at least one gene AMR target detected.

Table 34. Detection or Not of Antimicrobial Resistance for Patients in Table 19 with a

Bacterial Pathogen Found

Table 35. Antimicrobial Resistance Genes Details

Table 36: Sequences Detected for each AMR Gene Target for each Patient in Table 19 with a Bacterial Pathogen and at least one AMR Gene Detected

[0122] The three components of the panel refine both diagnosis and treatment options beyond what any one aspect of fee panel could provide. Specifically, the immune response scores distinguish between commensal and pathogenic presence of a microbe or microbes.

[0123] For example, detection of H. influenzae and/or P. aeruginosa by next- generation sequencing in an asthma patient could lead to either diagnosis of no infection and therefore no treatment if the panel doesn’t indicate that the presence of H. influenzae and/or P. aeruginosa is causing an immune response that is contributing to asthma symptoms, or diagnosis of infection and subsequent treatment with one of several antibiotics to which H. influenzae and/or P. aeruginosa is sensitive if the panel indicates that the infection is likely contributing to an immune response that drives asthma symptoms. Antibiotic treatment may be beneficial for these patients.

[0124] Additionally, information on the host immune response can refine the diagnosis of detection of the pathogen. If the immune response is indicative of response to the pathogen, then a diagnosis of infection may be made. If fee immune response is not indicative of infection by the pathogen, then a diagnosis of no infection (i.e., commensal presence of the microbial species) may be made. After diagnosis of infection rather than commensal presence of fee microbial species, the antimicrobial resistance information can refine fee available treatment options for a given patient.

[0125] For example, Table 19 demonstrates a correlation between some pathogen detection and immune response in many cases. For example, when a virus was not detected, the viral immune response score for that type of pathogen was also <1 in all eases. Similarly, the H. influenzae immune score was negative for samples that did not have H. influenzae detected, and the gram- negative bacterial immune response scores were <1 when no gram-negative bacterial pathogen was detected in all cases in Table 19.

[0126] However, in other cases, a lack of correlation between pathogen detection and immune response informs an appropriate diagnosis. For example, when comparing Patients 1 and 2 to Patients 3 and 4 in Table 19, all four patients had H. influenzae detected in sputum. However, Patients I and 2 had a gram negative bacterial immune response score >1.0, indicative of a gram- negative bacterial infection, and a score indicative for an immune response specific to H. influenzae (>1.0) while Patient 3 had neither immune response (scores less <1.0 for both gram- negative bacterial immune and HL influenzae response). Patient 4 had an immune response to a gram negative bacterial pathogen (score >1.0), but not an H. influenzae response (score <1). Therefore, a physician could use this information to make a diagnosis of an infection with H. influenzae in Patients I and 2, but not Patients 3 and 4. As shown in Table 19, in addition to H. influenzae. Patient 4 also had P. aeruginosa detected, a gram negative bacterial pathogen. Patient 4 did have an immune response to a gram negative bacterid pathogen (score >1). Therefore, a physician could use this information to make a diagnosis of infection with P. aeruginosa as opposed to HL influenzae for Patient 4. Further, after diagnosing each of Patients I and 4 with a gram-negative infection (Zf. influenzae and P. aeruginosa, respectively), observing antimicrobial resistance detection informs potential treatment options for these two patients. Specifically, as shown in Table 34, Patient 1 did not show expression of resistance genes to any of the antibiotics shown. However, Patient 4 showed resistance to the aminoglycoside and beta-lactam classes, both of which are active against gram negative bacterial pathogens and hence would have been potential treatment options for both patients. Based on this information from the panel, the physician may determine that those two classes are available for Patient I , but not indicated for Patient 4.

[0127] As described above, multiple genes are detected in order to determine detection of a pathogen. Observing expression of multiple genes that could potentially be detected results in greater sensitivity compared to choosing one target gene, as is often standard for qPCR. For instance, for//. influenzae (Table 20), among the patients who tested positive, there were different combinations of genes detected (hgpC, re I B, uppB, vapD, yafQ_l). In Patient I , the panel detected all genes except uppB. If the uppB gene had been selected for a qPCR assay, Patient 1 could be misdiagnosed as not having an //. influenzae infection. But it is clear by the detection of four other genes (hgpC, relB, vapD and yafQJ) that the pathogen is present Similarly, a misdiagnosis for C albicans could have been made for Patient 5 if another assay was used. As shown in Table 32, the panel showed no presence of the MEY_05088 gene in the patient sample. In a qPCR assay targeting the MEY_05088 gene only, the patient could be misdiagnosed as not having a C albicans infection. However, the panel indicated expression of three other C. albicans genes (MEY_03361 , MEY_04223, and MEY_06242), yielding a more accurate diagnosis.