TOOLS AND METHODS FOR GENETIC TESTS USING NEXT GENERATION

SEQUENCING

FIELD OF THE INVENTION

The present invention relates to tools and methods for use in genetic tests involving high performant sequencing techniques. More particularly, the invention relates to a robust robust multiplex PCR method wherein the respective primers for amplifying the different amplicons are physically isolated from one another. The invention further relates to quality control methods allowing a stringent quality control of genetic tests carried out according to the present invention.

BACKGROUND OF THE INVENTION

Mutations/variations in the human genome are involved in most diseases, going from monogenetic to multifactorial diseases, and acquired diseases such as cancer. Even the susceptibility to infectious diseases, and the response to pharmaceutical drugs, is affected by the composition of an individual's genome. Most genetic tests, which screen for such mutations/variations, require amplification of the DNA region under investigation. However, the size of the genomic DNA that can be amplified is rather limited. For example, the upper size limit of an amplified DNA fragment in a standard PCR reaction is about 2kb. This contrasts sharply with the total size of 3 billion nucleotides of which the human genome is build up. As more and more mutations/variations are found to be involved in disease, there is a need for robust assays in which different DNA regions, that harbor the different mutations/variations, are analyzed together. This may be achieved through (more complex) multiplex amplification reactions.

In this application, PCR is the method of amplification of DNA that is mainly described, however the embodiment of this invention may be used for any amplification technique known to the art, such as isothermal amplification (rolling circle amplification), ligase chain reaction, nucleic acid sequence-based amplification (NASBA), padlock probes, single strand displacement amplification, and whole genome amplification.

In a classical multiplex PCR reaction, different fragments are amplified in a single tube, simply by adding all pairs of amplicon-specific primers to a reaction mixture. The higher the number of primers that are combined in a single PCR reaction, the higher the chance that particular primer interactions (such as primer-dimerization), and aspecific primer/template

interactions occur, so that particular amplicons fail to amplify. There is thus a limitation in the number of amplicons that can be co-amplified when primers are simply mixed. Certain primer combinations work in one multiplex reaction, but not in another multiplex reaction. Also the amount of primers that are added may affect the success of a multiplex PCR amplification. The determination of amplicons that can be co-amplified, and the development of a robust multiplex PCR reaction of these amplicons, needs to be determined empirically by trial and error and is rather time-consuming. Each multiplex PCR reaction needs its own optimization. In a robust multiplex PCR reaction, 5-15 amplicons can be combined. A higher number of amplicons may be multiplexed, however at the expense of the robustness of the assay which affects the success rate of a multiplex amplification.

Robust multiplex PCR reactions of a large number of amplicons may be achieved when the different primers are physically restricted. Here we propose different methods to trap molecular components allowing a physically restricted amplification with minimal interference of other molecular components.

SUMMARY OF THE INVENTION

The new generation sequencing technology, such as parallel pyrosequencing in high-density picotiter reactors, allows the simultaneous sequencing of extreme high numbers of 25-500bp fragments and is currently mainly used for total genomic sequencing. The present invention is based on the finding that when using adapted sample preparation methods and protocols these sequencing technologies can also be used in other sequencing applications such as the resequencing of specific nucleic acid fragments in biological samples, as used in clinical diagnostics. In first instance it was found that a successful use of the said sequencing technologies preferably requires the availability of robust multiplex PCR reactions of a large number of amplicons. Therefore, it is a first object of the present invention to provide tools and methods for carrying out robust multiplex PCR reactions wherein the respective primers for amplifying the different amplicons are physically isolated from one another. This is achieved through the use of molecular complexes (hereinafter also referred to as DNA KNEX crystals) comprising two or more linked nucleic acids or analogues thereof wherein said molecular complexes comprises at least one forward primer and at least one reverse primer suitable for the PCR amplification of a given nucleic acid fragment. Preferably said molecular complexes are soluble in an aqueous solution.

In a preferred embodiment a molecular complex according to the present invention comprises one or more core molecules, which comprise a nucleic acid or a plurality of linked nucleic acids or analogues, for instance via avidin-streptavidin/biotin interactions. Said core

molecules further comprise at least two binding sites linking said core molecule to other core molecules and/or to functional molecules, which are nucleic acids or analogues thereof comprising either of the said PCR primers. In a more preferred embodiment the binding sites in said core molecule are two or more biotin molecules, while the functional molecules comprise at least one biotin molecule (Figure 1). This allows to firmly link the respective core and functional molecules using avidin or streptavidin as a linker molecule. The molecular complexes according to the present invention can also be used after immobilisation on a support, for instance on a support coated with avidin or streptavidin, which allows for binding the molecular complex through interaction with free biotin molecules in the molecular complex. Preferably, said support is the wall of a reaction vessel, such as a test tube.

In a second object the present invention provides a method for carying out PCR amplification of a DNA fragment said method comprising the use of molecular complexes according to the present invention wherein each of said complexes comprise one or more forward primers and one or more reverse primers allowing the amplification of said DNA fragment (Figure 2). More preferably said PCR method is a multiplex PCR for the simultaneous amplification of different DNA fragments wherein multiple sets of molecular complexes are used (Figure 3). The respective sets of molecular complexes each comprise complexes comprising one or more forward primers and one or more reverse primers specific for the amplification of one of the multiple DNA fragments to be amplified. The PCR method of the present invention is of particular interest for the amplification of a genomic DNA fragment or set of fragments for the purpose of sequencing a section of said genomic DNA.

New sequencing techniques allow a very large throughput and the generation of large amounts of data. They may even generate much more sequence data than needed for the analysis of one sample. In order to make full use of the capacity of a technique, and therefore at an economical cost, different samples need to be pooled in one single sequencing run. This requires that the origin of the DNA fragments derived from the respective samples can be traced. This can for instance be achieved by incorporating a same DNA tag in all DNA fragments derived from a given sample. The sequencing of these DNA fragments, including the incoporated tags, subsequently allows to analyze the parts of the fragments to be investigated as well as to determine the origin of the respective fragments. Therefore in a third object the present invention provides a method for introducing a DNA tag into a DNA fragment (Figure 4), including cDNA fragments, derived from a sample comprising genetic material said method comprising the steps of: i. the amplification of a first DNA fragment in a PCR-reaction using a first set of amplicon specific primers, each of said primers comprising an adaptor, preferably one adaptor for the forward primer and another adaptor for the reverse primer and

ii. subsequently amplifying said first DNA fragment in a PCR reaction using a second set of primers directed against the said adaptor sequences, wherein either or both of said second set of primers comprises a DNA-tag.

In a more preferred embodiment a same tag is introduced in more than one DNA fragment derived from a sample, said method comprising the steps of: i. the amplification of a first set of DNA fragments in a multiplex PCR-reaction using a first set of amplicon specific primers, wherein each forward primer of said set of amplicon specific primers comprises a same adaptor and wherein each reverse primer of said set of amplicon specific primers comprises a same adaptor and ii. subsequently amplifying said first set of DNA fragments in a PCR reaction using a second set of primers directed against the said adaptor sequences, wherein either or both of said second set of primers comprises a DNA-tag.

In a particular embodiment the first set of amplicon specific primers of step (i) are comprised in a molecular complex according to the first object of the present invention. In this case the amplification of the first set of DNA fragments involves the use of a method according to the second object of the present invention.

Preferably either or both of the second set of primers used in step (ii) comprises an additional adaptor for further processing purposes.

The introduced DNA tag is a known nucleic acid sequence, more preferably the DNA tag comprises one or more primary tags wherein each primary tag comprises a known sequence. Most preferably the DNA tag comprises a repitition of a same primary tag. When the relation between each of the introduced DNA tags and the samples from which the DNA fragments are derived is documented, it becomes possible to identify from which sample a DNA fragment is derived based on its sequence in as far a unique same DNA tag was introduced in all DNA fragments derived from a same sample. In a more preferred embodiment the information on the relation between each of the introduced DNA tags and the samples from which the DNA fragments are derived together with the determination of the sequences of said fragments allows to automatically report the results of the sequence analysis for each of said samples.

Finally it was found that the quality control of diagnostic procedures involving a biological sample, particularly when the analysis of the sample comprises the sequencing of nucleic acid fragments comprised in said sample, can be significantly improved by adding one or more nucleic acids comprising a known sequence to that sample prior to its analytic processing. Retrieval of said known sequences at the end of the analytical process will allow to verify the origin of said sample. Therefore, in fourth object the present invention provides a method for marking and subsequently verifying the origin and/or identity of a biological sample said method comprising the steps of:

i. providing an isolated biological sample; ii. adding to said sample one or more marker nucleic acids, each such marker nucleic acid comprising a known nucleic acid sequence, said known sequence being unique for each of said added marker nucleic acids; iii. documenting the relation between the identity and/or origin of said biological sample and the addition of said one or more marker nucleic acids to the sample; iv. detecting the presence in said biological sample the known nucleic acid sequences comprised in the marker nucleic acids; v. verifying whether the sequences detected in step (iv) are in accordance with the documentation obtained in step (iii) in order to verify the identity and/or origin of said biological sample.

Preferably, the presence in the biological sample of said known nucleic acid sequences is detected by PCR amplification of said sequences and subsequently sequencing the amplification products.

The method according to the present invention is particularly usefull in the quality control of the sequencing of nucleic acid fragments comprised in a biological sample. Preferably the known sequences comprised in the marker nucleic acids are processed in the same way as the DNA fragments of the sample. In a particular embodiment this implies that in the DNA fragments obtained by PCR amplification of the known nucleic sequences of the added marker nucleic acids and in the DNA fragments obtained by PCR amplification of the nucleic acid sequences to be characterized by sequencing a same DNA tag is simultaneously incorporated (Figure 6). The anticipated presence of the DNA tag in the DNA fragments derived from the sample and in the DNA fragments comprising the known sequences derived from the marker nucleic acids allows for a very stringent quality control of both the origin and the processing of the samples. In a more preferred embodiment the detection of the introduced DNA tag and that of the known nucleic acid sequences are used in the automatic reporting on the results of the analysis of the DNA fragments derived from the biological samples.

DETAILED DESCRIPTION OF THE INVENTION

Legends to the figures

Figure 1. KNEX-DNA-crystals. A. KNEX-DNA-crystal build up of biotinylated oligonucleotides, the core-oligonucleotide contains 3 biotin moieties, the functional oligonucleotides contain 1 biotin moiety. Here, two functional oligonucleotides contain a 3'

amplicon-specific domain which will allow amplification of the respective amplicon, and an adapter domain which will be used as priming site in another polymerization/amplification process. Streptavidin is used for cross-linking all oligonucleotides. B. Crosslinking of the core-oligonucleotides and streptavidin allows the formation of a complex core-network. C Crosslinking of the core-oligonucleotides and functional oligonucleotides allow the formation of amplicon-specific KNEX-DNA-crystals.

Figure 2. Principle of local amplification at a amplicon-specific KNEX-DNA-crystal. The amplicon-specific region of a functional oligonucleotide in the KNEX-DNA-crystal can hybridize with a target DNA molecule from a genomic DNA sample, which then primes a polymerization reaction. The original molecule will dissociate from the KNEX-DNA-crystal when denatured, but the newly synthesed fragment remains attached to the crystal. When the temperature is reduced, the newly synthesized fragment can hybridize with a neighboring oligonucleotide at the crystal. Target oligonucleotides that are attached to the KNEX-DNA- crystal have the highest probability for binding, which then in turn can start a polymerization reaction. This process can be repeated for several cycles. In this way, local specific amplification is achieved, with minimal interference of other oligonucleotides.

Figure 3. Different amplicon-specific KNEX-DNA-crystals, each containing two different functional oligonucleotides that allow amplification of one specific amplicon are prepared separately. Different amplicon-specific KNEX-DNA crystals can be pooled in one mixture so that the different types of amplicons are generated in an isolated means in the mixture.

Figure 4. Use of amplicon-specific KNEX-DNA-crystals in multiplex amplifications and which can be specifically tagged by a sample tag.

The different primers used in each step of the process have specific different melting temperatures, so that they can be specifically called in the process by their annealing temperature, and can therefore be mixed. a. For each amplicon, KNEX-DNA-crystals were build with the 2 respective primers for amplification of that amplicon, and different amplicon-specific KNEX-DNA-crystals were combined in the reaction. Each forward primer contains an amplicon-specific region (AS) and a universal forward adaptor (FA), analogously the reverse primers contain an AS region and another universal reverse adaptor (RA). The KNEX-PCR reaction favors amplification of the specific, versus aspecific fragments. All PCR fragments will then carry identical sequences at their ends. b. In a next amplification step, a sample tag (ST) is attached to each amplicon. By using a different primer pair per sample, all amplicons from one individual can be differently tagged from the amplicons from another individual. The forward and reverse primer may even contain a different sample tag (ST). This step also allows transfer of the amplicons (i.e. through the generation of their daughter amplicons) from a solid- phase status to a free status in solution, so that they can be easy manipulated in next steps. The primers are build up of three parts, the 3' end is complementary to the universal adapters attached in the previous step, allowing priming of all amplicons obtained in the previous step; a unique sample tag (ST), and a third adaptor (again different adaptors for the forward (A) and reverse primers (B)). The A and B adaptors will allow further manipulation, such as an water-in-oil-emulsion-PCR and pyrosequencing in a GS-FLX sequencing protocol.

Because of the use of sample-tagged fragments, all fragments from all individuals can now be processed simultaneously. After analysis, fragments can be traced to the original sample through the sample tag.

Figure 5. Generation of libraries of bar code molecules.

In first instance a recombinant vector is generated that contains an adapter sequence. A first primer against this adapter sequence will allow amplification of the molecular bar code region

in later steps. The adaptersequence is positioned so that PCR products of a reasonable length will be obtained in these later steps.

An oligonucleotide population, obtained in one oligosynthesis reaction, contains a 25- nucleotide long random tag sequence, flanked by two different restriction enzymes (RE). They also contain a DS adapter sequence, which is used as a priming site in order to generate double stranded oligonucleotides. The random sequence contains also an adapter sequence, which will be used for binding of the second primer that will allow amplification of the molecular bar code region in later steps.

Cloning of this oligonucleotide will generate 4 ²⁵ (1125899906842624) different types of bar code molecules. The format will allow generation of libraries without the need for immediate subcloning of all molecular bar codes. Subclones can be generated from this masterlibrary only when needed.

Figure 6. Use of molecular bar code molecules in multiplex parallel sequencing genetic tests.

The different primers used in each step of the process have specific different melting temperatures, so that they can be specifically called in the process by their annealing temperature, and can therefore be mixed.

Three bar code molecules are added to a sample to be analyzed, a differrent set for each sample. Then DNA is extracted, including the bar code molecules. a. Then a KNEX-PCR reaction, enabling local efficient independent amplifications, is performed. For each amplicon, KNEX-amplicon-specific DNA-crystals are prepared containing the 2 respective primers for amplification of that amplicon, and different KNEX- DNA-crystals are combined in the reaction. Each forward primer contains an amplicon- specific region (AS) and a universal forward adaptor (FA), analogously the reverse primers contain an AS region and another universal reverse adaptor (RA). The KNEX- PCR reaction favors amplification of the specific, versus aspecific fragments. All PCR fragments will then carry identical sequences at their ends. b. In this amplification step, a sample tag (ST) is attached to each amplicon. By using a different primer pairs per sample, all amplicons from one individual can be differently tagged from the amplicons from another individual. The forward and reverse primer may even contain a different sample tag (ST). This step also allows transfer of the amplicons (i.e. through the generation of their daughter amplicons) from a solid-phase status to a free status in solution, so that they can be easy manipulated in next steps. The primers are build up of three parts, the 3' end is complementary to the universal adapters attached in the previous step, allowing priming of all amplicons obtained in the previous step; a unique sample tag (ST), and a third adaptor (again different adaptors for the forward (A) and reverse primers (B). The A and B adaptors will allow further manipulation, such as an water-in-oil-emulsion-PCR and pyrosequencing in a GS-FLX sequencing protocol.

Because of the use of sample-tagged fragments, all fragments from all individuals can now be processed simultaneously. After analysis, fragments can be traced to the original sample through the sample tag. Moreover, since the sample tag will be also attached to the fragments derived from the bar code molecules, it can be verified if one is still dealing with the original sample, so that no sample switching has occurred. The bar code molecules and sample tag will also allow automaton of laboratory and reporting protocols.

Figure 7. visualisation by electrophoresis of DNA fragments amplified using DNA KNEX crystals in a multiplex PCR

Description

Robust multiplex PCR reactions of a large number of amplicons may be achieved when the different primers are physically restricted. Here we propose different methods to trap molecular components allowing a physically restricted amplification with minimal interference of other molecular components.

In a first method, for each DNA fragment that one wants to analyze, two primers that bind to this fragment and that allow amplification of that DNA fragment, are bound to beads, e.g. two biotinylated primers bound to streptavidin coated beads. In an analogous manner, beads are made for each DNA fragment that one wants to amplify, using specific primers to the respective DNA fragment. In the final multiplex amplification reaction, an aliquot of all these different beads mixtures is pooled. Although free primers may be still in solution, specific amplification will mainly occur at the beads, and primers bound to a bead will have the highest probability of priming the specfic PCR reactions at that bead, therefore allowing physical restriction of different amplification processes, thereby increasing the efficiency of the multiplex amplification. However, because of their gravity, beads will sediment to the bottom of the solution, thereby hampering the amplification. This can be overcome by restricting the beads themselves, e.g. by binding them to the wall of tubes, e.g. streptavidin- coated PCR tubes.

Another means would be the construction of light molecular structures, instead of beads, to which the primers are bound and which therefore remain in solution. Again here, they may be in turn fixed, e.g. to the wall of tubes. Such molecular structures might be obtained, as, what we call, molecular KNEX-DNA or molecular KNEX-DNA-crystals (see Figure 1). In KNEX-DNA structures, the two types of amplicon-specific primers are bound many times at different positions to another DNA molecule, and/or a network of smaller DNA molecules (e.g. oligonucleotides). KNEX-DNA crystals were chosen because small DNA molecules, and modifications thereof, are easily synthesized, modified and manipulated using well established technology at an economical price. This even applies to larger DNA molecules, such as plasmids. Indeed, by techniques such as nick translation, modified nucleotides can be incorporated allowing the formation of KNEX-DNA. The term DNA molecule includes linear oligomers of natural or modified monomers or linkages, e.g., deoxyribonucleosides, ribonucleosides, anomeric forms thereof, peptide nucleic acids (PNAs), and the like, the monomers are linked by phosphodiester bonds, or analogs thereof. The DNA molecules of the present invention can include non-natural nucleotide analogs. For the preparation of amplicon-specific crystals, other polymer forming molecules may be eventually used.

A KNEX-DNA-crystal used for PCR amplification of a given amplicon is called an amplicon- specific KNEX-DNA-crystal. In an amplicon-specific KNEX-DNA-crystal, the 2 amplicon- specific primers for amplification of a given amplicon in a PCR reaction are physically restricted and in close proximity, so that amplicons are amplified in/at their own crystal containing its own 2 amplicon-specific primers, with no, or minimal, interference of other primers in the solution. Therefore, isolated PCR reactors will be generated containing the 2 amplicon-specific primers in a tube, so that different amplicon-specific KNEX-DNA-crystals, which were prepared separately in a bulk solution, can be combined in a single multiplex PCR reaction.

These amplicon-specific primers may potentially contain an adaptor sequence, which is shared by all, or part of all, amplicon-specific primers, for cleavage and/or binding to universal primers in potential later steps of a process. A cleavable site might also be introduced by other means, such as a site in the oligonucleotide that can be cleaved by photons (UV-light). Another cleavable linker could be a disulfide, in which the cleavage occurs by a reducing agent, for example dithiothreitol (DTT), beta-mercaptoethanol, etc.

Since that each amplicon is amplified isolated at its own crystal-reactor, once that the amplification conditions for an individual amplicon have been optimized, there will be no, or minimal, need for empirical multiplex PCR optimization and design. Moreover, the number of amplicons that can be multiplexed may exceed 10-100 amplicons, so that this step can be easily transferred to any other gene, and even large or multiple genes.

The crosslinking of the components in the KNEX-DNA-crystal may be achieved by any manner known in the art. It can be either a covalent or non-covalent interaction. Covalent chemical crosslinking can be accomplished using standard coupling agents. For example, water-soluble carbodiimide can be used to link the 5'-phosphate of a DNA sequence to an amine-moiety through a phosphoamidate bond. Other linkage chemistries to join an oligonucleotide include the use of N-hydroxysuccinamide (NHS) and its derivatives, for example.

Other examples of interactions which can form the basis for crosslinking include interactions between biotin/streptavidin, biotin/avidin, biotin/biotin-binding-molecule (e.g. NEUTRAVIDINT modified avidin (Pierce Chemicals Rockford, IL), glutathione S- transferase(GST)/glutathione, antibody/antigen, antibody/antibody-binding-molecule, dioxigenin/anti-dioxigenin, DNP(2,4-dinitrophenyl)/anti-DNP antibodies, maltose-binding-

protein/maltose, chelation (e.g. (Co2+, Ni2+)/hexahistidine, pluronic coupling technologies, and so on.

A crystal might be even formed without the use of true separate junctional molecules, such as the use of oligonucleotides containing psoralen moieties, or even by Watson-Crick base pairing between oligonucleotides themselves. The latter may nevertheless even contain binding moieties, e.g. biotin, not for the purpose of forming the crystal but for adding functional groups to the crystal needed for later processing of a crystal. Depending of the type of crosslinking, some structures will be more stable and therefore more useful in certain applications. E.g. complexes that are made through Watson-Crick basepairing between oligonucleotides, and that are used for amplification, may decompose during a denaturation step at 94°C.

Specifically we describe here KNEX-DNA-crystals where the crosslinking is realized by biotin and streptavidin (Figures 1 and 2).

The DNA-core molecule of a KNEX-DNA-crystal can be 1 'core'-oligonucleotide that carries at least two biotin groups, at its 5' end, its 3' end, and/or internal sites. To a solution containing 'core'-oligonucleotide, streptavidin is added, as well as the two amplicon-specific primers are added. Sub-solutions may first be prepared of some of these components, but a some stage they end up complexing in the same solution. Indeed, the order and amount by which the different components are added might affect the composition. Certain compositions may be better suited for certain amplifcations. For example, addition of the 'core'-oligonucleotide and streptavidin may form molecular type of beads, and subsequent addition of the amplicon specific primers may cover the surface of the molecular beads. Each streptavidin molecule is able to bind four biotin groups, and therefore a three-dimensional network or KNEX-DNA-crystal is formed between the three oligonucleotides. The density of each type of oligonucleotide in the KNEX-DNA-crystals can be controlled by the ratio of 'core'-oligonucleotide to the 2 amplicon-specific primers, as well as by the concentration of streptavidin. Even the ratio of the 2 amplicon-specific primers may be controlled. After crystallization, free biotin might be added to saturate the still free binding sites on streptavidin molecules, so that different amplicon-specific KNEX-DNA-crystals will not aggregate when mixed. If certain amplicons remain difficult to be amplified in the multiplex PCR, different KNEX-DNA-crystals may be evaluated. For example, KNEX-DNA-crystals in which the 'core'- oligonucleotide contains varying numbers of internal biotin groups, so that the local concentration of particular amplicon-specific primers is further increased. By keeping incubation times during an amplification step limited, diffusion of products may be even more

restricted. Possibly one or both primers should be still freely available to some extent in the PCR reaction in order to obtain the most efficient amplification process. Example 1 shows a proof of principle experiment.

Rather than an oligonucleotide, larger DNA molecules can be used as core-DNA molecules in KNEX-DNA crystals. For example, plasmid DNA, either linearized or not, can be biotinylated by nick translation to varying extents depending on the relative concentration of biotinylated nucleotide building blocks.

Each core of the KNEX-DNA-crystal might be 1 DNA molecule, such as an oligonucleotide with at least 2 biotin groups; or an aggregate of different DNA molecules, such as oligonucleotides with at least 2 biotin group that are polymerized, linked by streptavidin molecules or base-pairing. Each of these latter DNA molecules preferentially contains at least two different biotin groups and different types of such oligonucleotides might be used (i.e. oligonucleotides having different sequence and/or modifications). This is a more economical and easier way to synthesize large core-molecules, and allows the synthesis of a KNEX-DNA-crystal were amplicon-specific primers are more physically spread over the crystal.

The preparation of such structures may not be 100% efficient, so that all components are not incorporated in the KNEX-DNA-crystals and may be present as individual components in a solution. It is also possible that during later processing of the crystals, part of the crystals will degrade partly or complete, possibly only temporally, to their individual components of which the crystals were built up. During processing of a solution containing these crystals, as well as the free components, these free components may interfere in later processing. Their contribution may improve or worsen the actual processing of the crystal, or have no effect at all.

In a multiplex PCR amplification of amplicon-specific KNEX-DNA-crystals, different amplicon-specific KNEX-DNA-crystals will be mixed. For each amplicon-specific KNEX-DNA- crystal, two amplicon-specific primers will be used, so that all these different amplicon- specific KNEX-DNA-crystals have to be synthesized separately, and should be mixed afterwards (Figure 3). This should be, once optimized, be done only once, and therefore in bulk and with minimal effort.

The presence of a biotin at the 3' end of the core-DNA molecule (e.g. oligonucleotide) will not allow priming or interference in a PCR reaction. The sequence in this 'core'-

oligonucleotide can be also used for specific functions, such as binding to other oligonucleotides (e.g. for capturing). Also it may contain recognition sites for restriction enzymes, so that after binding to a complementary oligonucleotide, the KNEX-DNA-crystal can be cleaved after the amplification step (e.g. for disruption of the KNEX-DNA-crystal, so that amplified material becomes more accessible in later steps of a protocol). The latter might be also achieved by a cleavable linker (e.g. by UV) in the 'core'-oligonucleotide, which would be a more universal means of cleavage than the use of restriction enzymes, since a particular recognition site for a restriction enzyme might be present in the amplified amplicon.

In KNEX-DNA-crystals, the 2 amplicon-specific primers are possibly positioned in a 3- dimensional structure. This may affect the efficiency of amplification. When the 2 amplicon- specific primers are bound to beads they are only positioned on a 2 dimensional surface. Since KNEX-DNA-crystals have a relatively low molecular weight, they remain homogenously solubilized in the reaction mixture and are readily accessible to reagents.

By reducing the specific incubation times and/or temperatures to a strict minimum, one will limit the extent of diffusion of products that are generated in a process, and in this way promote local processing of products.

KNEX-DNA-crystals may be even build up by 2 to 4, or even more, biotinylated 'core'- oligonucleotides in a more programmed manner. The formation of the crystals, i.e. the binding of the 'core'-oligonucleotides through biotin-streptavidin bounds, may be preceded by self-mediated DNA assembly of these 'core'-oligonucleotides. This self-mediated DNA assembly may allow the formation of rather specifically well-designed KNEX-DNA-crystals.

One could overcome the preparation of amplicon-specific structures (amplicon- specific beads, amplicon-specific KNEX-DNA crystals) if adaptors are added to the amplicon specific primers. In that case, only one type of KNEX-DNA structures containing only the two primers directed to the adaptor sequences needed to be prepared. Preferentially, the forward amplicon-specific primers contain a specific type of adaptor, while the reverse amplicon- specific primers contain another specific type of adaptor.

Alternatively, a standard multiplex amplification might be developed in a more robust way using solely amplicon specific primers that also contain an adaptor, preferentially a different adaptor for the forward primer than the reverse primers. However, all forward amplicon- specific primers carry the same adaptor, while all reverse amplicon-specific primer carry the same other adaptor. First, all amplicon-specific primers (with their adaptors) are added

together and a classical PCR reaction is performed for a few cycles. In classical multiplex PCR reactions, the aspecific amplifications and the inefficient amplification of specific fragments increases with the number of cycles. If only a few cycles are performed, the specific fragments are still present in a relative high proportion, and if they proceed in a second more specific PCR step, with only 2 primers directed against the adaptor sequences of the first primers, they are more favorably amplified. Moreover, since all fragments then have a common priming site, there will be less primer-dependent preferential amplification that is observed in a classical multiplex PCR reactions. Moreover, if amplicons are designed in such a way that they have a similar length, there will be also no amplicon-length dependent specific amplification. One could also further enrich the specific amplification fragments from the aspecific fragments, through binding of the PCR products obtained after the first or second round of PCR to a molecule, or molecules, that recognize the desired fragments.

One could also perform an enrichment of the target DNA to be sequenced before the amplification step, which would reduce aspecific binding of primers during PCR. For example, in case that one wants to amplify the different exons of a given gene, one could hybridize total genomic DNA, either fragmented or not (e.g. by a (rare-cutter) restriction enzyme) to (a) DNA fragment(s) that contain sequences of this gene itself. For example, this could be a plasmid containing a cDNA derived from a transcript from this gene. In that case, they will bind all homologous genomic DNA fragments at the respective site of the coding sequences. This DNA fragment used for capturing cannot be a template in the subsequent PCR reaction, since all primers are located to intronic regions. The plasmid itself could be biotinylated by nick translation, which than can be bound to streptavidin-coated tubes, or streptavidin-coated beads (e.g. streptavidin-coated magnetic beads), and in this way would be an easy means in further separation steps.

At this moment, most genetic tests only analyse the coding region of a gene, i.e. the exons and their exon/intron junctions. Such tests will benefit from robust multiplex amplifications. However, in instances were also complete introns need to be analysed, covering the whole gene in amplicons in a multiplex PCR reactions may be still too challenging, given the size that some genes can obtain. Here, a strategy could be to enrich the complete gene through hybridization of total DNA (genomic DNA, mitochondrial DNA,...), either fragmented or not (e.g. by a (rare-cutter) restriction enzyme) to (a) DNA fragment(s) that contain sequences of this gene itself. For example, this could be again a plasmid containing a cDNA derived from a transcript from this gene. These plasmids will then bind all targeted DNA fragments from a genomic sample at exonic regions. If the DNA was cut with a rare-cutter restriction enzyme, also intronic sequences will be captured together with the

exon sequences. The plasmid itself could be biotinylated by nick translation, which than can be bound to streptavidin-coated tubes, or streptavidin-coated beads (e.g. streptavidin-coated magnetic beads), and in this way would be an easy means for isolation of the relevant fragments. The isolated fragments could then be amplified by PCR using random primers. Preferentially the random primers contain adaptor sequences for further processing in sequencing protocols, either after size-fractionation or not. On the other hand, the enriched DNA fragments can feed in standard sequencing protocols, such as the generation of a DNA-library. This enrichment step could be also performed after the generation of a DNA- library in a standard sequencing protocol, rather then before the generation of the DNA library. The latter protocol would have the added advantage that only sequences of the DNA sample under investigation will be sequenced, and not the plasmid that was used for isolating/enriching the target DNA. Indeed, these plasmids could in the end possibly 'contaminate' the sample under investigation. Since the adaptorsequences that are added to the different DNA molecules of a DNA library only allow further processing in the sequencing protocol, plasmids used for enrichment will not contain these adaptorsequences so that they cannot be processed or sequenced furher.

This strategy could be even applied for more than 1 gene, i.e. different plasmids directed two the different genes under investigation that are combined (e.g. all genes involved in a given pathway or disease, all genes belonging to a certain class of proteins). Analogously, one could even enrich complete genomic regions, e.g. through the use of BAC- clones that recognise the region to be sequenced and therefore provide a means for enrichment of a subgenomic region.

An analogous strategy could be also used for enrichment of RNA (either converted to cDNA or not), than DNA.

New techniques allow the generation of large amounts of data, e.g. sequence data. They may even generate much more sequence than needed for the analysis of one sample. In order to make full use of the capacity of a technique, and therefore at an economical cost, different samples need to be mixed in one single experiment, in which each sample is tagged differently. The characterization of the tag in the final analysis allows the identification of the sample of which the fragment was derived, while the characterization of the attached fragment will allow analysis of the fragment under investigation of that sample.

Addition of a DNA-tag to a DNA fragment of a sample that one wants to analyze can be obtained in an amplification process in which at least one amplicon-specific primer contains the tag as an adaptor. However, this will be still very costly. E.g. when one wants to analyze 30 amplicons (60 primers) in one sample, and one wants to combine 200 samples,

one needs 12000 primers if a tag is included in both amplicon-specific primers, and 6030 primers if a tag is included in one of the two amplifcon-specific primers.

A much lower number of primers will be needed if tagging is performed in a two step protocol using adaptors, which is illustrated here (Figure 4). In the first PCR reaction, amplicon-specific primers with an adaptor are used. All forward amplicon-specific primers contain the same adaptor, all reverse amplicon-specific primers contain the same other adaptor. In a second PCR reaction, primers are used directed against these adaptor sequences. They all include a unique tag sequence, and possibly additional adaptor sequences for further processing. In such a strategy, only 460 primers are needed if two tags per amplicon are included, or 260 primers if only one tag is included.

Moreover, the adaptor tag primers can be used for any amplicon, irrespective of the genomic region, or gene, from which it was derived, if all first step PCR reactions are performed with adaptor-amplicon specific primers that contain the same 'universal' forward and reverse adaptors.

Through careful design of the oligo's one could mix all primers in one reaction, so that only certain primers participate in a process at a given moment. E.g. in a two-step PCR assay, in which the oligonucleotides that are used for priming in the first amplification step have a higher meltingtemperature than the oligonucleotides that are needed in the second PCR process, one could add all primers together in one tube; if the PCR profile is first programmed for the first cycles to use an annealing temperature that uses a melting temperature of the first set of oligo's, only these oligo's can act preferentially as primer; if in later cycles the annealing temperature is lowered, the other primers are able to prime the amplification. Of course, the first primer pairs are still able to prime reactions, however if they were initially added at a much lower concentration, the second set of oligonucleotides will be kinetically favored.

For example, tags composed of 5 nucleotides will allow 1024 unique different tags. It should be noted that nucleotide incorporation errors can be introduced in an oligosynthesis reaction. Especially in assays in which single molecules are analyzed, such synthesis errors may not be tolerated. Indeed, a synthesis error will change the identity of the tag, and the attached fragment will be incorrectly correlated with the wrong sample. This may be overcome by repeating the tag one or several times in the tag-oligonucleotide. For example, in a tag-oligonucleotides in which the 5-nucleotide tag is repeated 4 times, it will be unlikely that the same synthesis/processing error occurs in each of the 4 tags. Identification of the sample will then be done on the basis of the tag that is most observed in a 4 5-tag sequence (e.g. 3 out of 4 tags are identical, while the fourth tag has one nucleotide difference). The

repetition of a tag will provide a means to detect, and possibly correct, the misincorporation erros of oligo's through synthesis. This repetition strategy might me combined by other strategies aiming to correct for incorporations, such as error-correcting barcodes (1).

The origin of each amplicon, i.e. the DNA of the individual from which the PCR fragment was generated, can thus be determined through characterization of the tag region, either found at the 5'-end and/or 3'-end, e.g. by sequencing. The remainder sequence of the amplicon will allow the analysis of the sample.

At the end of a sequence reaction, a 3'-tag sequence may be difficult to determine. Indeed, lower resolutions of sequencing signals are obtained the more one progresses in the sequencing reaction, or because of the increase in background signal the more one progresses in the sequencing reaction. Loss of synchronization during the sequencing reaction results in an increasing background so that the actual sequence cannot be further read-out the more the sequencing reaction progresses. E.g., each step in a sequencing reaction will not be 100% efficient, so that a small fraction of an 'amplicon-clone' is not elongated, but will be elongated in the next step of the sequencing reaction, and will thus result in background signal. These background signals will accumulate as the sequencing reaction progresses. Therefore, a 3'-tag might be difficult to determine. This problem could be overcome by reducing the length of the amplicons to be sequenced, however this might not be an option since less sequence information will be obtained per amplicon. The use of poly-homo-stretch-tags could overcome this problem. In a poly-homo-stretch-tag, each letter of the code is not one nucleotide, but a stretch of 2 or more identical nucleotides. The sequence of the tag will therefore generate more intense signals, e.g. in a pyrosequencing reaction, so that it can be more easy discriminated from the background, and therefore allow that most sequence of the amplicon is used for characterization of the fragment under investigation.

Any process is prone to errors, especially at moments when contents of tubes are transferred to other tubes. Especially in diagnostics this cannot be tolerated.

One way to overcome this is by molecular bar coding of the sample itself, i.e. in the sample, as is illustrated here. In this way the whole process is quality assured from the moment the molecular barcode is added. Preferably that molecular bar code is added to the sample as early as possible. The earliest possible moment would be when the specimen for analysis is collected, e.g. the molecular bar code is already present in the collector tube (e.g. collecting tube for a blood sample).

The molecular bar code should be found at the end of the complete process, and this guarantees quality assurance.

Molecular bar coding is feasible in assays that analyze individual molecules and that analyze a large number of molecules so that a small proportion of all tested single molecules can be used for identification of the molecular bar code and therefore monitor the sample.

A molecular bar code could be a unique DNA fragment, for each sample under investigation one unique DNA fragment. This unique DNA fragment is flanked with adaptors, so that it can processed together in the same strategy as all other fragments under investigation, so that the unique DNA barcode fragment proceeds through the whole process in an analogous strategy as the actual fragments of the sample under investigation. The molecular bar code therefore needs the same features as the molecules under investigation, so that they can be processed simultaneously. It might be a synthetic molecule, or a plasmid, or even a recombinant bacteria containing such a plasmid. During the process they may be free in solution (e.g. blood) or bound to certain components of the sample solution (e.g. blood cells)

A 5 nucleotide molecular bar code allows 1024 different unique molecular bar code molecules. A 10 nucleotide molecular bar code allows about 1000000 unique different molecular bar code molecules, and so on. The synthesis of such a high number molecular bar code molecules is, however, quiet costly.

A more economic favorable way would be the use of more than 1 molecular bar code molecule to mark a sample, e.g. 3 molecular bar code molecules, all of them having the same adaptors to allow processing with the samples under investigation (Figure 5). In combination with a sample tag as described above, when 3 molecular bar codes are used, one would only need 30 molecular bar code molecules to obtained 1000 unique different combinations, compared to 1000 bar code molecules that would be needed if only one bar code molecule was used per sample. The combination with a sample tag will provide an extra control of the process, since the correct combination of molecular bar code and sample tag should be found at the end of the process which ensures that complete process is quality-assured (Figure 6). Mixing of three type of molecular bar code molecules, in which the molecular bar code is coded by 5 nucleotides, i.e. 3072 (1024*3) molecular bar code molecules, allows more than 10 ⁹ (1024 ³ ) different combinations of unique molecular bar codes, and allows many strategies of tagging (e.g. sample specific, sample-specific and lab- specific, ...). The molecular bar coding format can be used for different types of parallel sequencing techniques.

Such molecular bar code molecules can be easily generated by mutagenesis using olignucleotides that contain randon nucleotide stretches. The repertoire of molecular bar code molecules could be even expanded if random sequences are inserted at two positions, rather than at one position in a vector library.

Apart from the qualilty-assucrance, such molecular bar codes allow automation of the laboratory and reporting protocols.

Not only can these molecular bar codes be used in a DNA assay, in which the molecular bar code and sample are characterized in the same assay. Molecular DNA bar codes may be also used for any assay, in which the molecular DNA bar code typing and actual assay or two different assays (e.g. determination of protein in doping tests, and so on). In this way any sample, especially solution samples, can be labeled. Moreover, the labeling can be completely anonymous, so that willing incorrect manipulation can be hardily done.

Example 1 : Amplicon-specϊfic KNEX-DNA-crystal multiplex amplification of fragments that cannot be co-amplified in a multiplex PCR reaction.

If KNEX-DNA-crystals facilitate amplification in a multiplex amplification through keeping the amplicon-specific primers physically restricted, one would expect that amplicons which cannot be co-amplified in a standard multiplex PCR reaction should be co-amplified in amplicon-specific KNEX-DNA-crystal multiplex PCR reactions. If so, this could indicate a proof-of-principle of an amplicon-specific KNEX-DNA-crystal multiplex PCR assay.

In the past it was found that a 538bp-amplicon containing exon 1 (Figure 7) (lane 7), and a 341bp-amplicon containing exon 3 (lane 8), of the SCNN1A gene, cannot be co- amplified in a standard multiplex PCR reaction. In the standard co-amplification reaction, the 341bp-amplicon fails to amplify when both are combined (lane 9). In amplicon-specific KNEX-DNA-crystal multiplex PCR, in which amplicon-specific KNEX-DNA-crystals were build using 5 pmol of a double-biotinylated, 34 nucleotide long oligonucleotide (5'-B- CCGπAACCCGATATCGGCCCGGGCCTπAAACC-B-3 ¹ (biotin groups are underlined); in the presence of 2.5 pmol (lanes 1 and 4), 5 pmol (lanes 2 and 5), and 7.5 pmol (lanes 3 and 6) amplicon-specific primers; either in the presence of 1.25 pmol (lanes 1-3) or 2.5 pmol (lanes 4-6) streptavidin, the 341-bp-amplicon is amplified to variable degrees, of which the best result is obtained in lane 4. The fact that still individual amplicons can be visualized by electrophoresis, instead of DNA-smears because of aggregates of KNEX-DNA-crystals, indicates that not all amplicon-specific primers are aggregated in KNEX-DNA-crystals in this

experiment or that streptavidin-biotin bounds were broken during the process. Lane D contains a 10Obp DNA ladder.

Example 2: Sequencing

Parallel pyrosequencing in high-density picotiter reactors by the Genome Sequencer (454 Life Sciences Inc, Roche Applied Science) allows the determination of millions of nucleotides in a single run hours (2).

In such sequencing assays, total genomic DNA is randomly fragmented to 25-500bp fragments, depending on the sequencing technology, to which adapters are ligated. Amplification is then performed with two universal primers against these adaptors so that all fragments are amplified simultaneously. For resequencing purposes of a gene, this 'universal' amplification step with two primers cannot be performed, since each amplicon needs to be amplified by its own specific pair of two primers. Moreover, since the coding region of genes can be more than 5kb, many amplicons need to be amplified.

Multiplex amplification of all these amplicons in a single, or a limited number of, PCR reaction(s) would thus render amplicon-sequencing more convenient and cheaper. This may be achieved by amplicon-specific KNEX-DNA-crystal multiplex PCR as described above. In combination with sample tags, different samples can be pooled (Figure 4), and in combination with molecular bar codes the complete process can be quality-assured (Figure 6).

Example 3: Determination of large genomic deletions, insertions or duplications

Some mutations might result in complete loss, duplication or insertion of larger DNA fragments. Such fragments may encompass 1 or more complete exons of a gene. Such mutations are missed in mutation scanning assays (e.g. sequencing) of the complete coding region, and the exon/intron junctions, of a gene. Indeed, such assays make use of PCR products obtained in the plateau phase of a PCR reaction. In case of an individual, having a recessive disease and who is heterozygous for a complete deletion of 1 exon, the corresponding region in the. other gene will still be amplified, and in case that the mutation in that other gene is located outside that exon, a normal sequence will be read. It will then be concluded that the individual carries the wild type nucleotides on both genes, while in fact the sequence was read from 1 gene only. Techniques such as MLPA (multiplex ligation probe mediated amplification) or QMPSF (Quantitative Multiplex PCR of Short fluorescent

Fragments) allow the detection of the complete loss or duplication of exons. These quantitative techniques are based on multiplex amplification and therefore have their limitations in robustness and accuracy. Since that these assays are based on quantitative principles, preferential amplification of smaller fragments may lead to false conclusions or results that cannot be analyzed. Amplicon-specific KNEX-DNA-crystal multiplex PCR amplifications may thus result in more robust quantitative multiplex amplification, MLPA, or QMPSF assays. In case that amplicon-specific KNEX-DNA-crystal multiplex PCR amplification is quantitative, such deletions or insertions may be even detected in sequencing technologies that perform parallel sequencing starting from single template molecules. Indeed, in these sequencing technologies each amplicon is sequenced at a certain coverage or redundancy. If a higher or lower amount of sequenced amplicons are detected in a multiplex of amplicons, the deletion or duplication of that amplicon may then be concluded. Especially since most deletions/duplications will be covered by more than 1 amplicon, different amplicons are in fact controls for each other. These parallel sequencing assays may then also provide quantitative information besides qualitative sequence determination. Analogously, more complex copy number variations (CNVs) can be determined.

Definitions:

* 'core'-molecules and 'functional -molecules: a 'core'-molecul in a KNEX-DNA-crystal is a molecule arm which essentially forms the frame of the KNEX-DNA-crystal, a 'functional'-molecule in a KNEX-DNA-crystal is a molecule which will be used in a particular process and that is attached to the frame of the KNEX-DNA-crystal. It is possible that a 'core'-molecule and 'functional'-molecule may be attached at a same junction. It is possible that the 'functional'-molecule has also 'core'-molecule properties, and vice versa.

References

Hamady, M., Walker, J.J., Harris, J.K., Gold, NJ. and Knight, R. (2008) Error- correcting barcoded primers for pyrosequencing hundreds of samples in multiplex. Nat Methods, 5, 235-7.

Margulies, M., Egholm, M., Altman, W.E., Attiya, S., Bader, J.S., Bemben, L.A., Berka, J., Braverman, M.S., Chen, Y.J., Chen, Z. et al. (2005) Genome sequencing in microfabricated high-density picolitre reactors. Nature, 437, 376-80.