Some content of this application is unavailable at the moment.
If this situation persist, please contact us atFeedback&Contact
1. (WO2018140695) CONSTRUCTION OF NEXT GENERATION SEQUENCING (NGS) LIBRARIES USING COMPETITIVE STRAND DISPLACEMENT
Note: Text based on automatic Optical Character Recognition processes. Please use the PDF version for legal matters

CONSTRUCTION OF NEXT GENERATION SEQUENCING (NGS) LIBRARIES USING COMPETITIVE STRAND DISPLACEMENT

FIELD OF THE INVENTION

[0001] The invention pertains to construction of next-generation DNA sequencing (NGS) libraries for whole genome sequencing, whole exome sequencing, targeted resequencing, sequencing-based screening assays, metagenomics, or any other application requiring sample preparation for NGS.

BACKGROUND OF THE INVENTION

[0002] Next Generation Sequencing (NGS) has evolved into a very powerful tool in molecular biology, allowing for the rapid progress in fields such as genomic

identification, genetic testing, drug discovery, and disease diagnosis. As this technology continues to advance, the volume of nucleic acids that can be sequenced at one time is increasing. This allows researchers to sequence larger samples, as well as to increase the number of reads per sample, enabling the detection of small sequence variations within that sample.

[0003] As the volume and complexity of NGS processing increases, so does the rate of experimental error. While much of this error occurs in the sequencing and processing steps, they can also occur during the sample preparation steps. This is particularly true during the conversion of the sample into a readable NGS library by which adaptor sequences are attached to the ends of each fragment of a fragmented sample (library fragment) in a uniform fashion.

[0004] There are several types of errors that can occur during the execution of next generation sequencing (NGS), and it is important to be able to differentiate between true rare variants, such as rare alleles or mutations that exist in the patient and errors that arise from sequencing and/or sample preparation. Particularly problematic are errors that are introduced during library construction, prior to library amplification via polymerase chain reaction (PCR). Such errors can propagate during PCR, leading to multiple copies of sequences containing the error, making it difficult to distinguish between the errors and true variants. The general strategy used to overcome this is consensus calling, whereby sequence reads that are PCR copies of a single, original fragment are grouped together and compared to similar groups of copies, derived from other original

fragments, which overlap in sequence. If a variation is present in one group of clones and not the others then it is most likely an error propagated by PCR, whereas variations present in several groups are most likely true variants. In order to perform this analysis one must be able to differentiate between clones derived from one molecule and those derived from another.

[0005] The terms "fragments", "target fragments", or "inserts", as used herein, refer to fragments of DNA, created from the fragmentation of a DNA sample, which are processed into an NGS library and sequenced. The processing of these fragments usually involves end repair and A-tailing, followed by the addition of sequencing adapters and amplification.

10006] The terms "depth of coverage", "coverage depth" or "target coverage", as used herein, refer to the number of sequenced DNA fragments (i.e., a reads) that map to a genomic target The deeper the coverage of a target region (i.e., the more times the region is sequenced), the greater the reliability and sensitivity of the sequencing assay. In general, a coverage depth of 500— 1000X, or higher, is often required for the detection of low frequency sequence variations.

[0007] The terms "adenylated", or "pre-adenylated", as used herein, refer to a state by which a strand of DNA has an adenosine 5 '-monophosphate (AMP) covalently attached to its 5' -terminal phosphate via a pyrophosphate bond. The terms "adenylate", or "adenylation", as used herein, refer to the process of covalently attaching an AMP either to a protein side chain or to the 5 '-terminal phosphate of a DNA strand. The term "adenyl group", as used herein, refers to an AMP that is either covalently attached to, or transferred between, a protein sidechain and/or DNA strand.

[0008] The term "consensus sequence", as used herein, refers to a sequence obtained by comparing multiple sequences within a family of sequences. Sequence variations that are present in some, but not in the majority of sequences, in the family may be designated as errors and subsequently removed from the analysis. On the other hand, sequence variations that are present in the majority of sequences within a family may be designated as true variants that were present in the original genetic material being analyzed. The term "consensus calling", as used herein, refers to the process to determining if a genetic variation is a true variation or an error.

[0009] The term "variant calling", as used herein, refers to the process of determining if a sequence variation is a true variant derived from the original sample, and thus used in the analysis, or the result of a processing error and thrown out.

[0010] The term "family", as used herein, refers to a group of reads that are determined to be duplicates based on their having the same start stop sites and/or UMIs. In variant calling, large families with multiple clones are desirable since they can be used to build stronger consensus sequences than those with only a few clones to compare. For very small family sizes with one or two clones, a consensus cannot be called, resulting in potentially important data being thrown out.

[0011] The term "deduplication", or "dedup", as used herein, refers to the removal of reads that are determined to be duplicates, from the analysis. Reads are determined to be duplicates if they share the same start stop sites and/or UMI sequences. One purpose of deduplication is to create a consensus sequence whereby those duplicates that contain errors are removed from the analysis. Another purpose of deduplication is to estimate the complexity of the library. A library's "complexity", or "size", as used herein, refers to the number of individual sequence reads that represent unique, original fragments and that map to the sequence being analyzed.

[0012] The terms "start stop sites", or "fragment ends", as used herein, refer to the sequences at the 5' and 3' ends of a sheared library fragment that become directly ligated to the sequencing adapters. Start stop sites can be used to determine if two similar sequences are derived from separate molecules or are cloned copies of the same original fragment. In order for different original fragments to have the same start stop sites, the shearing events that created them would have had to cleave at exactly the same sites, which has a low probability. Clones, on the other hand, should always have the same start stop sites. As such, any fragments that share the same start stop site (due to random shearing), are usually considered duplicates. The term "position-based", as used herein, refers to the use of stop start sites as a criterion for determining whether or not a read is a duplicate of another.

[0013] A "start stop collision", as defined herein, is the occurrence of multiple unique fragments that contain the same start stop sites. Due to the rarity of start stop collisions, they are usually only observed when either performing ultra deep sequencing with a very high number of reads, such as when performing low variant detection, or when working with DNA samples that have a small size distribution, such as plasma DNA. As such, start stop sites may not be enough in those scenarios since one would run the risk erroneously removing unique fragments, mistaken as duplicates, during the deduplication step. In these cases, the incorporation of UMIs into the workflow can potentially rescue a lot of complexity.

[0014] The term "UMI", or "Unique Molecular Identifier", as used herein, refers to a tag, consisting of a sequence of degenerate or varying bases, which is used to label original molecules in a sheared nucleic acid sample. In theory, due to the extremely large number of different UMI sequences that can be generated, no two original fragments should have the same UMI sequence. As such, UMIs can be used to determine if two similar sequence reads are each derived from a different, original fragment or if they are simply duplicates, created during PCR amplification of the library, which were derived from the same original fragment.

[0015] UMIs are especially useful, when used in combination with start stop sites, for consensus calling of rare sequence variants. For example, if two fragments have the same start and stop site but have a different UMI sequences, what would otherwise have been considered two clones arising from the same original fragment could now be properly designated as unique molecules. As such, the use of UMIs combined with start stop sites often leads to a jump in the coverage number since unique fragments that would have been labeled as duplicates using start stop sites alone will be labelled as unique from each other due to them having different UMIs. It also helps improve the Positive Predictive Value ("PPV) by removing false positives. There is currently a lot of demand for UMIs, as there are some rare variants that can only be found via consensus calling using UMIs.

[0016] "PPV", or Positive Predictive Value, is the probability that a sequence called as unique is actually unique. PPV = true positive / (true positive + false positive).

"Sensitivity" is the probability that a sequence that is unique will be called as unique. Sensitivity = true positive / (true positive + false negative).

[0017] Two errors that occur during library construction, and which are reduced by the present invention, are the formation of (1) fragment chimeras and (2) adaptor dimers.

[0018] Fragment chimeras are the result of library fragments ligating with one another without the adaptor sequences, resulting in longer fragments that contain unrelated sequences juxtaposed to one another. These unrelated sequences would thus be mistakenly read as a continuous sequence. As such, suppression of fragment chimera formation during library construction is important for reducing downstream sequencing errors.

[0019] Adapter dimers are the result of self-ligation of the adapters without a library insert sequence. These dimers form clusters very efficiently, reduce reaction efficiencies, and consume valuable space on the flow cell. This is especially problematic when

dealing with ultra-low DNA input quantities in the picogram range. At such low DNA input levels, adapter dimers can constitute a majority of the NGS library molecules formed, thus reducing the amount of useful information generated by DNA sequencing. For this reason, suppression of adapter dimer formation during library construction is a very important but challenging task.

[0020] Provided herein are high throughput methods for NGS library construction based on novel adapter ligation strategies that can minimize the formations of both fragment chimeras and adaptor dimers and accurately convert DNA samples into sequencing libraries in under a day. These and other advantages of the invention, as well as additional inventive features, will be apparent from the description of the invention provided herein.

SUMMARY OF THE INVENTION

[0021] The invention pertains to construction of next-generation DNA sequencing (NGS) libraries for whole genome sequencing, targeted resequencing, sequencing-based screening assays, metagenomics, or any other application requiring sample preparation for NGS. The proposed methods each consist of a two-step ligation process by which a first sequencing adapter is ligated to end repaired DNA fragments via blunt end ligation and a second sequencing adapter is then ligated to the first ligation product via splint end ligation. This process of NGS library preparation will be referred to here as Competitive Strand Displacement (CSD). Although initial work has focused on attachment of P5 and P7 adaptors for Illumina sequencing, this method could be used on alternate platforms which also require the attachment of one or more synthetic sequences (Ion Torrent® sequencing platform for example).

[0022] In a first embodiment of the method (Figure 1), fragmented DNA is subjected to an end-repair reaction producing blunt 5 ' phosphorylated inserts with free 3' OH ends. This may be accomplished with T4 Polynucleotide Kinase (PNK) and T4 DNA polymerase, or any other combination of enzymes that leaves blunt ends with 5' phosphates and 3' hydroxyl groups. Following end-repair, the first sequencing adaptor (P5 or P7 for Illurnina platforms) is attached to the 3' end of insert DNA via blunt ligation using T4 DNA ligase; one strand of the adaptor is 5' phosphorylated to facilitate ligation, while the complementary strand is blocked on the 3' end with a dideoxy nucleotide (ddN) to prevent ligation. The second sequencing adaptor is then attached to the 5' ends of biological inserts through a splinted ligation reaction linking the 3' ends of adaptor molecules to the phosphorylated 5' ends of the inserts. This ligation can be performed using Taq DNA ligase, or any other ligase capable of performing splinted ligations with little activity on blunt-ended substrates (Ampligase, 9°N, Tth, etc.)- Since these ligases prefer splinted ligation, adaptor dimers are minimized, which mitigates the need for size-selection post-ligation. Following the second ligation, the newly constructed library molecules can either be sequenced directly ("PCR-free") or amplified via PCR prior to sequencing. Representative examples of the first and second strands of the first sequencing adapter for the first embodiment are provided as SEQ ID NOs:3-10, and SEQ ID NO: 17, respectively. Representative examples of the second sequencing adapter for the first embodiment are provided as SEQ ID NOs:l-2. All representative examples for the first embodiment are offered to illustrate, but not to limit, the claimed invention.

[0023] In a second embodiment of the method described above, a mutant T4 DNA ligase, K159S (see U.S. Application Serial No. 15/426,543, referenced in its entirety), is used for the first ligation (Figure 2). This mutant cannot utilize ATP to adenylate substrates prior to ligation, and is thus only capable of ligating substrates that were pre-adenylated. This feature can be exploited by performing ligation with pre-adenylated sequencing adaptors as this will only result in adaptor-to-insert ligation events (rather than insert-to-insert), which greatly suppresses chimera formation. Furthermore, the ligation efficiency of wild-type T4 DNA ligase is thought be hindered by "aborted ligation" events where adenylated ligase units transfer adenyl groups to inserts, but fail to effectively join 5' and 3' ends. In such instances, ligase units will be quickly re-adenylated rendering them inactive on DNA that has already been adenylated. Since the mutant cannot be adenylated, aborted ligation events are not problematic and ligation efficiency is increased relative to that of wild-type T4 DNA ligase. Representative examples of the first sequencing adapter for the second embodiment are provided as SEQ ID NOs:l 1-16. All representative examples for the second embodiment are offered to illustrate, but not to limit, the claimed invention.

[0024] In a third embodiment of the present invention, the pre-adenylated adaptors in the first ligation step are ligated onto the 3' ends of the target fragments via a wild type T4 ligase, instead of the K159S mutant, and in the absence of ATP, thus preventing the formation of fragment chimeras (Figure 3).

|0025] In a fourth embodiment of the methods described above, sequencing adaptors can be ligated to the 5 ' end of inserts first (Figure 4). In this embodiment, the 3 ' end of the first sequencing adaptor (PS or P7 for Illumina sequencing) is attached to the 5' end of phosphorylated inserts via blunt ligation. To prevent dimer formation, this adaptor is not 5' phosphorylated in the double-stranded portion, so it will not ligate to inserts or other adaptor molecules. The second adaptor sequence (P5 or P7 for Illumina sequencing) is attached to the 3' end of inserts via splinted ligation using a single-stranded oligo that has phosphate groups on the 5' ends and C3 spacers on the 3' ends. This ligation can be performed using Taq DNA ligase, or any other ligase capable of performing splinted ligations with little activity on blunt-ended substrates (Ampligase, 9°N, Tth, etc.). Since the ligase prefers splinted ligation, adaptor dimers are minimized, which mitigates the need for size-selection post-ligation.

[0026] In a fifth embodiment of the present invention, there is an RNA base on the 3' end of the truncated ligation helper oligonucleotide, instead of a ddN. In this case, both the 5' end of the adaptor and the 3' end of the truncated stem are ligated to the insert. However, the stem is then cleaved off via the activity of an RNase H2 enzyme which cleaves 5' of the RNA base. After an SPRI cleanup step, the second ligation takes place (Figure 5).

[0027] In a sixth embodiment of the present invention, the first sequencing adapter has a tag sequence on its 5' end which serves to independently label the sense and antisense strands of the target on their 3' ends (Figures 6 and 7). These sequence tags are not limited to any particular length or sequence. The bases can be degenerate, fixed, or a combination of both. Modified bases can also be used. As before, the 5' end of the adapter is pre-adenylated and ligated on to the repaired 3' end of the target by either the Kl 59S mutant T4 DNA ligase, or via a wildtype T4 ligase in the absence of ATP. The second sequencing adaptor is then annealed to the ligated first adapter at its

complementary stem sequence, leaving a gap which spans the tag sequence. The gap is then filled in with a polymerase, creating an in situ UMI which is complementary to the first UMI. After the fill in step, the 3 ' end of the newly created in situ UMI is ligated on to the 5' end of the target fragment. Optionally, this is followed by a PCR amplification step using primers that prime off of the first and second adapter sequences and may optionally add additional sequences such as sample barcodes and/or P5/P7 sequences. Representative examples of the first strand of the first sequencing adapter for the sixth embodiment are provided as SEQ ID NOs: 18-33. Representative examples of the second strand of the first sequencing adapter for the sixth embodiment are provided as SEQ ID NOs:34-49. A representative example of the second sequencing adapter for the sixth embodiment is provided as SEQ ID NO:50. Representative examples of optional forward and reverse primers that could be used for the PCR amplification step of the sixth embodiment are provided as SEQ ID NOs:75-98, and SEQ ID NOs:51-74, respectively. These particular representative forward and reverse PCR primers contain P5 and P7 adapter sequences, respectively, as well as sample barcode sequences. All representative examples for the sixth embodiment are offered to illustrate, but not to limit, the claimed invention.

[0028] The seventh embodiment of the present invention is a variation of the sixth embodiment in which the second sequencing adapter has an additional sequence that is complementary to tag sequence added during the first ligation step (Figure 7). As a result, no gap is present after the second sequencing adapter anneals to the first ligation product and the fill-in step with a DNA polymerase is unnecessary prior to ligation. This can be accomplished by using a finite number of variable tags. For example, this embodiment could potentially consist of a plurality of first sequencing adapters, each with one of 24 distinct variable tag sequences, and a plurality of second sequencing adapters, each with one of 24 distinct sequences which are complementary to the variable sequences of the first sequencing adapters.

[0029] In any of the above embodiments, unique molecular identifiers (UMIs) and sample barcodes can be incorporated into one or both of the sequencing adaptors.

Molecular identifiers can be constructed using fixed or degenerate sequences of any length compatible with Illumina sequencers.

[0030] In any of the above embodiments, one or more of the sequencing adapters used for the first and/or second ligations are shortened versions of the full sequencing adapters, in which case the remaining parts of the sequencing adapters are added later via PCR with tailed primers

[0031] The invention can be used for any application involving DNA sequencing, but is especially valuable for cancer diagnostics where detection of rare variants in mixed populations of tumor and normal DNA is crucial. The invention can also be used to construct sequencing libraries from Formalin-Fixed Paraffin-Embedded (FFPE) samples. The invention can also be used to construct sequencing libraries from ultra-low inputs of DNA with or without PCR, which may aid in forensic or microbiological studies where limited quantities of DNA are available and/or PCR cannot be tolerated.

[0032] Unlike the prior art which requires size-selection due to formation of adaptor dimers, the invention features a ligation strategy that does not require size selection. Lack of size-selection enables superior recovery of DNA, which greatly increases library complexity/coverage and sensitivity to low frequency variants. Adaptor dimers are also problematic for library quantification and sequencing, because standard methods of DNA quantification are greatly skewed by their presence. This can cause suboptimal cluster density and significantly reduce the number of reads aligning to actual samples, which increases sequencing costs. Also, unlike the prior art, the embodiment of the invention employing K159S does not create chimeras via ligation, which should greatly improve detection of rare structural variants associated with cancer.

DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 illustrates the first embodiment of the Competitive Strand

Displacement (CSD) method. The first step consists of the attachment of a first sequencing adapter (2) to the DNA target fragment (1) via blunt end ligation catalyzed by T4 DNA ligase. Said first sequencing adapter consists of a first and second DNA strand. The first DNA strand (4) has a C3 blocking group on its 3' end, a phosphate group on its 5' end (5' PO), and consists of a first sequence (6) which is complementary to, but longer than, the second DNA strand (3) and a second, non-complementary tag sequence (5) that contains the first sequencing primer binding site and, optionally, a UMI and/or sample barcode sequence. The second DNA strand (3) is a truncated

oligonucleotide, with a dideoxy nucleotide base (ddN) at its 3' end, and serves to facilitate the blunt end ligation of the 5' PO of the second DNA strand to the 3' OH of the target fragment, leading to the first ligation product (7). The second step consists of the attachment of a second sequencing adapter (8) to the first ligation product (7) via splint end ligation catalyzed by Taq ligase. Said second sequencing adapter has a 3' OH and consists of a first sequence (10) that is complimentary to the first sequence (6) of the first sequencing adapter and a second sequence (9) that contains the second sequencing primer binding site and, optionally, a second UMI and/or sample barcode sequence. Since the length of the complementary sequence of the second adapter (10) is longer than that of the truncated oligo of the first adapter (3), the second adapter is able to displace the truncated oligo during the annealing step that precedes the splint end ligation. The splint end ligation leads to the final library product (11).

[0034] Figure 2 illustrates the second embodiment of the C3D method. The elements of the second embodiment are similar to those of the first embodiment, the difference being that the first sequencing adapter is pre-adenylated at the 5' end (5' ppA) of the first strand and that the blunt end ligation is catalyzed by a mutant T4 DNA ligase, K159S, that cannot use ATP as a substrate for ligation and can thus only ligate the pre-adenylated strand of the first adapter to the 3' OH of the target fragment.

[00351 Figure 3 illustrates the third embodiment of the CSD method. The elements of the third embodiment are similar to those of the second embodiment, the difference being that the blunt end ligation is catalyzed by wildtype T4 DNA ligase in the absence of ATP. Since ATP is unavailable as a substrate for ligation, the wildtype T4 DNA ligase can thus only ligate the pre-adenylated strand of the first adapter to the 3 ' OH of the target fragment.

[0036] Figure 4 illustrates the fourth embodiment of the CSD method. The first step consists of the attachment of a first sequencing adapter (12) to the DNA target fragment via blunt end ligation catalyzed by T4 DNA ligase. Said first sequencing adapter consists of a first and second DNA strand. The first DNA strand has a 3' OH group on its 3' end and consists of a first sequence (14) which is complementary to, but longer than, the second DNA strand (IS) and a second, non-complementary tag sequence (13) that contains the first sequencing primer binding site and, optionally, a UMI and/or sample barcode sequence. The second DNA strand (15) is a truncated oligonucleotide with a C3 blocking group at its 3' end, a dephosphorylated 5' end, and serves to facilitate the blunt end ligation of the 5' PO of the target fragment with the 3' OH of the first strand of the first sequencing adapter, leading to the first ligation product (16). The second step consists of the attachment of a second sequencing adapter (17) to the first ligation product via splint end ligation catalyzed by Taq ligase. Said second sequencing adapter has a 5' PO and consists of a first sequence (40) that is complimentary to the first sequence (14) of the first sequencing adapter and a second sequence (41) that contains the second sequencing primer binding site and, optionally, a second UMI and/or sample barcode sequence. Since the length of the complementary sequence of the second adapter (40) is longer than that of the truncated oligo of the first adapter (IS), the second adapter is able to displace the truncated oligo during the annealing step that precedes the splint end ligation. The splint end ligation leads to the final library product.

[0037] Figure 5 illustrates the fifth embodiment of the CSD method. The elements of the fifth embodiment are similar to those of the second embodiment, the difference being that the truncated second strand (18) of the first sequencing adapter has an RNA residue at its 3' end. Said first sequencing adapter is then attached to the DNA target fragment via blunt end ligation catalyzed by the K.159S mutant T4 DNA ligase. Unlike the previous embodiments, both the first and second strands of the first sequencing adapter are ligated to the target fragment, the truncated second strand being ligated, via its 3'R, to the 5' PO of the target fragment, resulting in a first ligation product (19). The truncated second strands are then removed via RNase H2 cleavage, which occurs at the phosphodiester bond on the 5' side of an RNA residue. The resulting product (20) is similar to the first ligation product of the previous embodiments, the difference being that it has 3' RNA residues. Said 3' RNA residues are then ligated to the 5' PO ends of the second sequencing adapters during the second ligation step, resulting in a library product (21) with internal RNA residues.

[0038] Figure 6 illustrates the sixth embodiment of the CSD method. The elements of the sixth embodiment are similar to those of the second embodiment, albeit with the following differences. In this embodiment, the first DNA strand (24) of the first sequencing adapter (22) contains a variable tag sequence (26, 27) on its 5' end This serves to differentially label the sense and antisense strands of the target fragments during the first ligation step, leading to a first ligation product with each strand labeled differently (28). As with the previous embodiments, blunt end ligation is enhanced using a blocked and truncated second strand (23) that, in this embodiment, is complementary to the variable region (26,27) and part of the constant region (25) of the first DNA strand (24). During the second ligation step, the second sequencing adapter (29) anneals to the first ligation product via its sequence (30) that is complementary to the constant sequence added by the first sequencing adapter (25), but not to the variable region (26, 27). This results in a gap that is filled in with a DNA polymerase and a DNA ligase (31), leading to a final library product (32) with its sense and antisense strands labelled differently.

[0039] Figure 7 illustrates the seventh embodiment of the CSD method. The elements of the seventh embodiment are similar to those of the sixth embodiment, the difference being that the second sequencing adapter (33) has an additional sequence (34, 35) that is complementary to the variable tag sequence (36, 37) added by the first

sequencing adapter. As a result, no gap is created after the second sequencing adapter anneals to the first ligation product and no polymerase step is needed.

[0040] Figure 8A. Depth of coverage values for each of three replicate libraries, obtained using the method described in Example 1, are plotted for CSD (dark gray circles) and NEB (light gray circles) for 10 ng (left side) and 1 ng (right side) of DNA input. For the 10 ng DNA input, the average depth of coverage for CSD was 1009X, vs 598X for NEB. For the 1 ng DNA input, the average depth of coverage for CSD was 13 IX, vs 53X for NEB.

[0041] Figure 8B. Depth of coverage values for each of three replicate libraries, obtained from the experiment described in Example 1, are plotted for CSD (dark gray circles) and Kapa (light gray circles). The average depth of coverage for CSD was 1006X, vs 628X for Kapa.

[0042] Figure 8C. Depth of coverage values for each of three replicate libraries, obtained from the experiment described in Example 2, are plotted for CSD (dark gray circles) and NEB (light gray circles) for libraries derived from the "true" (left side) and "mock" (right side) cfDNA. For the "true" DNA input, the average depth of coverage for CSD was 276X, vs 77X for NEB. For the "mock" DNA input, the average depth of coverage for CSD was 241X, vs 104X for NEB.

[0043] Figure 8D. Depth of coverage values for each of three replicate libraries, obtained from the experiment described in Example 3, are plotted for CSD (dark gray) and NEB (light gray) for 1 ng (left side), 5 ng (middle) and 10 ng (right side) of DNA input. When compared with the NEB method, the average depth of coverage for CSD was 1.8X, 1.4X, and 1.3X higher with the lng, 5ng, or lOng of the FFPE derived genomic DNA, respectively.

[0044] Figure 9. Percent chimera values for each of three replicate libraries obtained from the experiment described in Example 4, are plotted for CSD (dark gray) and NEB (light gray) for libraries derived from the "true" (left side) and "mock" (right side) cfDNA. When compared with the NEB method, the average % of chimeras present for CSD was 1.6X lower with the "true" cfDNA input and 1.8X lower with the "mock" cfDNA input.

[0045] Figure 10A. Traces, generated with a Bioanalyzer DNA 1000 chip, that show the size distribution of DNA molecules present in each of three replicate libraries generated with the NEB or CSD methods from the sample DNA with 1% or 0.5% minor

allele fractions. The absence of dimer peaks at the 1 SO bp mark (39) for the CSD method, and presence of such peaks in the NEB method (38), demonstrates the reduced occurrence or adapter dimers for libraries prepared with CSD, when compared to those prepared with the NEB method.

[0046] Figure 10B. Traces generated with the Bioanalyzer DNA1000 chip (post-PCR) for each of three replicate libraries created with lOng of high quality genomic DNA sheared to 150bp, 200bp, or 300bp (gDNA extracted from cell line NA12878 procured from ATCC). For all three fragment lengths, there was an absence of dimer peaks that are typically observed in the 125bp-l SObp range.

[0047] Figure IOC. Traces generated with the Bioanalyzer DNA1000 chip (post-PCR) for each of three replicate libraries created with lOng or lng of high quality genomic DNA sheared to 200bp. For both input amounts, there was an absence of dimer peaks that are typically observed in the 125bp-150bp range.

[0048] The following examples illustrate, but do not limit the claimed invention.

EXAMPLE 1

[0049] This example demonstrates the enhanced depth of coverage obtained from NGS libraries, prepared from high quality genomic DNA, using the second embodiment of the CSD method as compared to that obtained when using either the NEB® Ultra™ II library (New England BioLabs) or Kapa Hyper Prep (Kapa Biosystems) methods. The high quality genomic DNA was extracted from cell line NA 12878 (ATCC). Either 1 or 10 ng of the extracted DNA was sheared to an average size of 150 bp using ultrasonic fragmentation (Covaris S220) and then subjected to end-repair, which included phosphorylation of the 5' ends with T4 Polynucleotide Kinase (PNK), for 30 minutes, followed by purification via 2.5X AMPure beads. For the CSD treatment, P7 adapters (SEQ ID NOs:l 1-16), hybridized to truncated, 3' ddN blocked oligonucleotides (SEQ ID NO: 17), were ligated onto the end repaired target fragments via blunt end ligation using the mutant Kl 59S T4 DNA ligase for 15 minutes, followed by a 15 minute heat kill step. P5 adapters (SEQ ID NO: 1 or SEQ ID NO:2) were then ligated onto the first ligation product using Taq DNA ligase for IS minutes, followed by purification using 2.SX AMPure beads. For the NGS treatment, libraries were prepared as per manufacturer's instructions. Both libraries were then subjected to a PCR-amplification with primers that contain sequences that are complimentary to the PS and P7 adapters under the following conditions: 98°C for 45 seconds, 12 cycles of: 98°C 15s, 60°C for 30 seconds, 72°C for 30 seconds, 72°C for 1 minute, 4°C hold. The libraries then underwent hybrid capture, using a custom panel of around 800 IDT Lockdown probes, to pull down fragments containing target sequences that were used to determine the depth of coverage values. The resulting, target enriched, product was purified via 1.8X AMPure beads and sequenced on a MiSeq® sequencer (Ilium ina) using 2 X 150 paired-end reads and following the manufacturer's protocol. The libraries were prepared in triplicate. Depth of coverage values for each of the three libraries obtained from CSD for 10 and 1 ng of DNA input, are plotted in comparison to those values obtained from the NEB (Figure 8A) and Kapa (Figure 8B) methods. When compared with the NEB method, the average depth of coverage for CSD was 1.7X higher withlO ng of DNA input, and 2.5X higher with 1 ng of DNA input (Figure 8A). When compared with the Kapa method, with 10 ng of DNA input, the average depth of coverage for CSD was 1.6X higher (Figure 8B). Depth of coverage values were determined by the number of unique reads (not counting PCR duplicates) that mapped to the expected target sequences that were enriched for via the 800 probe lockdown panel.

EXAMPLE 2

[0050] This example demonstrates the enhanced depth of coverage obtained from NGS libraries, prepared from circulating cell free DNA (cfDNA), using the second embodiment of the CSD method as compared to that obtained when using the NEBNext® Ultra™ II library. "True" cfDNA samples are real cell-free DNA isolated by Biochain from healthy individuals, while "mock" cfDNA samples are cell-line genomic DNA (NA12878) sheared to 150bp using a Covaris S2. Libraries were prepared with lng of the cfDNA using the CSD and NEB methods, as described in Example 1, in triplicate. When compared with the NEB method, the average depth of coverage for CSD was 3.6X higher with the "true" cfDNA input, and 2.3X higher with the "mock" cfDNA input (Figure 8C).

EXAMPLE 3

[0051] This example demonstrates the enhanced depth of coverage obtained from NGS libraries, prepared from low quality genomic DNA extracted from FFPE samples,

using the second embodiment of the CSD method as compared to that obtained when using either the NEB Ultra II library. The FFPE samples were procured from Asterand Bioscience. Libraries were prepared as described above using lng, 5ng, or lOng of the FFPE derived genomic DNA, sheared to an average size of 200 bp, as starting material. When compared with the NEB method, the average depth of coverage for CSD was 1.8X, 1.4X, and 1.3X higher with the lng, 5ng, or lOng of the FFPE derived genomic DNA, respectively (Figure 8D).

EXAMPLE 4

[0052] This example demonstrates the reduced chimera rate in NGS libraries prepared from cfDNA using the second embodiment of the CSD method as compared to that present in cfDNA libraries prepared using the NEB method. Libraries were prepared as described above, using lng of "true" or "mock" cfDNA as input, in triplicate. When compared with the NEB method, the average % of chimeras present for CSD was 1.6X lower with the "true" cfDNA input and 1.8X lower with the "mock" cfDNA input (Figure 9). The % chimera values were calculated based on the number of unique reads that were improperly aligned with the reference sequence (hgl9). Fragments categorized as "chimeric" have either (1) paired reads that face the same direction (same orientation), (2) paired reads that align to regions of the reference sequence that are greater than 3kb apart, and/or (3) paired reads that align to different chromosomes.

EXAMPLE 5

[0053) This example demonstrates the reduced occurrence of adapter dimers in NGS libraries prepared from high quality genomic DNA when using the second embodiment of the CSD method as compared to that present in libraries prepared using the NEB method. The high quality genomic DNA samples were extracted from two cell-lines, NA12878 and NA24385, and mixed at two different ratios, resulting it two mixtures having 1% and 0.5% minor allele fractions, respectively. Samples were sheared to 300 bp fragments. NEB libraries were created using a 0.9X AMPure ratio post-ligation, which is meant to size select away adaptor-dimer. CSD libraries were created with a 2.5X ratio post-ligation, which is too high to effectively remove full length adaptor dimers. NEB libraries were treated with 1.OX AMPure post-PCR to remove any residual

dimer, while CSD libraries were treated with a 1.8X ratio. The final library products were analyzed on a Bioanalyzer DNA1000 chip, by which traces were generated that showed the size distribution of DNA molecules present in each library. The absence of dimer peaks at the 150 bp mark for the CSD method without size selection indicates dimer formation is negligible or non-existent in libraries prepared with the CSD method (Figure 10A). Libraries prepared using the NEB method, on the other hand, still contain small amounts of adaptor dimer, despite two size selection steps, as is indicated by the small peaks at the 150 bp mark in the traces (Figure 10A).

EXAMPLE 6

[0054] This example demonstrates that the reduced presence of adapter dimers in NGS libraries prepared using the second embodiment of the CSD method is independent of the lengths of the target fragments used as the starting point. Libraries were created as described above with lOng of high quality genomic DNA, extracted from cell line NA12878, and sheared to 150bp, 200bp, or 300bp. As described above, the final library products were analyzed on a Bioanalyzer DNA1000 chip, generating size distribution traces. For all three fragment lengths, there was an absence of dimer peaks that are typically observed in the 125bp-150bp range (Figure 10B).

EXAMPLE 7

[0055] This example demonstrates that the reduced presence of adapter dimers in NGS libraries prepared using the second embodiment of the CSD method is independent of the amount of input DNA used as starting material. Libraries were created as described above with lOng or lng of high quality genomic DNA, extracted from cell line NA12878, and sheared to 200bp. For both input amounts, there was an absence of dimer peaks that are typically observed in the 125bp-150bp range (Figure IOC). For reference, the secondary peak at about 1500 bp merged with the upper marker is a known phenomenon due to over-amplification during PCR.

EXAMPLE 8

[0056] This example demonstrates the enhanced sensitivity achieved in NGS libraries prepared from high quality genomic DNA using the sixth embodiment of the CSD method as compared to that obtained when using the Kapa Hyper Prep method. The high quality genomic DNA was extracted from cell-lines NA12878 and NA24385 and mixed at a ratio of 1/100, generating a homozygous and heterozygous minor allele frequency of 1% and 0.5%, respectively. The genomic mixtures, with inputs ranging from 1 to 25 ng, were sheared to an average size of 150 bp using ultrasonic

fragmentation (Covaris S220), and then subjected to end-repair, which included phosphorylation of the 5' ends with T4 Polynucleotide Kinase (PNK), for 30 minutes, followed by purification via 2.5X AMPure beads. For the CSD treatment, truncated P7 adapters (SEQ ID NOs: 18-33), hybridized to truncated, 3' ddN blocked oligonucleotides (SEQ ID NOs: 34-49), were ligated onto the end repaired target fragments via blunt end ligation using the mutant K159S T4 DNA ligase for 15 minutes, followed by a 15 minute heat kill step. Truncated P5 adapters (SEQ ID NO:50) were then annealed to the constant sequence added by the first sequencing adapter (25 in Figure 6), but not to the variable region (26 and 27 in Figure 6). The resulting gap was filled in using Taq DNA polymerase, followed by ligation with Taq DNA ligase. This was followed by purification using 2.5X AMPure beads. The product was then subjected to a PCR-amplification with tailed primers containing the remaining portions of the P7 and P5 adapter sequences. P7 tailed primer sequences are listed as SEQ ID NOs:51-74 while P5 tailed sequences are listed as SEQ ID Nos:75-98. PCR conditions were as follows: 98°C for 45 seconds; 12 cycles of 98°C for 15s, 60°C for 30 seconds, 72°C for 30 seconds; 72°C for 1 minute; 4°C hold. For the Kapa treatment, libraries were prepared as per manufacturer's instructions. The libraries underwent hybrid capture, using a ~100kb custom panel of IDT Lockdown® probes, to pull down subsets of the mixed genotypes. In these subsets, there were 291 known nucleotide differences between the NA12878 and NA24385 sequences and these were used to assess the sensitivities and PPVs of the two library prep methods. The libraries underwent ultra-deep sequencing on a MiSeq® sequencer (Illumina) using 2 X 150 paired-end reads and following the manufacturer's protocol. This was followed by variant calling, using VarDict software. While there were three false positives called with the Kapa library at 20 ng input, there were zero false positives called with the Kapa library at 10 ng DNA input and the CSD library at both 10 and 20 ng DNA input, resulting in a PPV of one for both libraries. The number of false negatives, however, was 3x lower when using the CSD libraries at 20 ng input and 2x lower at 10 ng input, when compared to those gotten using the Kapa libraries at the same amounts of input The results are shown in Table 2.

Table 1: Sequences





Table 2: Sensitivity and Positive Predictive Values for Variant Calls using CSD or Kapa Prepared Libraries

[0057] All references, including publications, patent applications, and patents, cited herein are hereby incorporated by reference to the same extent as if each reference were individually and specifically indicated to be incorporated by reference and were set forth in its entirety herein.

[0058] The use of the terms "a" and "an" and "the" and similar referents in the context of describing the invention (especially in the context of the following claims) are to be construed to cover both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. The terms "comprising," "having,"

"including," and "containing" are to be construed as open-ended terms (i.e., meaning "including, but not limited to,") unless otherwise noted. Recitation of ranges of values herein are merely intended to serve as a shorthand method of referring individually to each separate value falling within the range, unless otherwise indicated herein, and each separate value is incorporated into the specification as if it were individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., "such as") provided herein, is intended merely to better illuminate the invention and does not pose a limitation on the scope of the invention unless otherwise claimed. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the invention.

[0059] Preferred embodiments of this invention are described herein, including the best mode known to the inventors for carrying out the invention. Variations of those preferred embodiments may become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventors expect skilled artisans to employ such variations as appropriate, and the inventors intend for the invention to be practiced otherwise than as specifically described herein. Accordingly, this invention includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by the invention unless otherwise indicated herein or otherwise clearly contradicted by context.