Methods of lowering the error rate of massively parallel DNA sequencing using duplex consensus sequencing
View Patent ↗Next Generation DNA sequencing promises to revolutionize clinical medicine and basic research. However, while this technology has the capacity to generate hundreds of billions of nucleotides of DNA sequence in a single experiment, the error rate of approximately 1% results in hundreds of millions of sequencing mistakes. These scattered errors can be tolerated in some applications but become extremely problematic when “deep sequencing” genetically heterogeneous mixtures, such as tumors or mixed microbial populations. To overcome limitations in sequencing accuracy, a method Duplex Consensus Sequencing (DCS) is provided. This approach greatly reduces errors by independently tagging and sequencing each of the two strands of a DNA duplex. As the two strands are complementary, true mutations are found at the same position in both strands. In contrast, PCR or sequencing errors will result in errors in only one strand. This method uniquely capitalizes on the redundant information stored in double-stranded DNA, thus overcoming technical limitations of prior methods utilizing data from only one of the two strands.
1. A method for detecting double-stranded deoxyribonucleic acid (DNA) molecules in a biological sample, comprising:
(a) tagging said double-stranded DNA molecules in said biological sample with a set of duplex tags, wherein said set of duplex tags comprises a plurality of different tag sequences, wherein each duplex tag of said set of duplex tags differently tags complementary strands of a double-stranded DNA molecule of said double-stranded DNA molecules in said biological sample to provide tagged strands, and wherein said tagging is performed with an excess of duplex tags as compared to said double-stranded DNA molecules;
(b) for each genetic locus in a set of one or more genetic loci in a reference genome, selectively enriching said tagged strands for subset of said tagged strands that map to said genetic locus, to provide enriched tagged strands;
(c) sequencing at least a portion of said enriched tagged strands to generate a plurality of raw sequence reads from said biological sample;
(d) grouping said plurality of raw sequence reads into a plurality of families, each family comprising raw sequence reads generated from a same parent polynucleotide, which grouping is based on at least one of (i) tag sequences associated with said parent polynucleotides and (ii) information from beginning and/or end portions of said raw sequences of said parent polynucleotides;
(e) collapsing said plurality of raw sequence reads grouped into said plurality of families into a plurality of consensus sequence reads, each consensus sequence read of said plurality of consensus sequence reads (i) comprising a plurality of consensus bases for each genetic locus in said set of one or more genetic loci and (ii) being representative of single strands of said double-stranded DNA molecules;
(f) for each genetic locus in said set of one or more genetic loci, quantifying said enriched tagged strands that map to said genetic locus for which complementary strands are detected in said plurality of consensus sequence reads; and
(g) for each genetic locus in said set of one more genetic loci, quantifying said enriched tagged strands that map to said genetic locus for which only one strand among complementary strands is detected in said plurality of consensus sequence reads, thereby detecting said double-stranded DNA molecules in said biological sample.
2. The method of claim 1 , further comprising quantifying the recovery of original parent polynucleotides from the biological sample that map to said genetic locus, wherein said recovery is determined at least in part on (i) said quantification of enriched tagged strands that map to said genetic locus for which complementary strands are detected and (ii) said quantification enriched tagged strands that map to said genetic locus for which only one strand among complementary strands is detected.
3. The method of claim 1 , further comprising sorting consensus sequence reads into paired sequence reads and unpaired sequence reads, wherein (i) each paired read corresponds to consensus sequence reads generated from a first tagged strand and a second differently tagged complementary strand derived from a double-stranded DNA molecule in said biological sample, and (ii) each unpaired read represents a first tagged strand having no second differently tagged complementary strand derived from a double stranded DNA molecule represented among said consensus sequence reads in said set of consensus sequence reads.
4. The method of claim 3 , further comprising quantifying at least two of (i) said paired sequence reads that map to each of said set of one or more genetic loci, (ii) said unpaired sequence reads that map to each of said set of one or more genetic loci, (iii) read depth of said paired sequence reads, and (iv) read depth of said unpaired sequence reads.
5. The method of claim 4 , further comprising determining a number of distinct double-stranded DNA molecules in said biological sample that map to a particular genetic locus in the set of one or more genetic loci based on the quantification of at least two of (i) said paired sequence reads that map to each of said set of one or more genetic loci, (ii) said unpaired sequence reads that map to each of said set of one or more genetic loci, (iii) read depth of said paired sequence reads, and (iv) read depth of said unpaired sequence reads.
6. The method of claim 4 , further comprising determining a number of distinct double-stranded DNA molecules in the biological sample that map to a particular genetic locus in the set of one or more genetic loci by quantifying said paired sequence reads and said unpaired sequence reads that map to the particular genetic locus.
7. The method of claim 1 , wherein said duplex tags are not sequencing adapters.
8. The method of claim 1 , wherein collapsing said plurality of raw sequence reads comprises collapsing raw sequence reads produced from amplified products of an original polynucleotide molecule in said biological sample back to said original polynucleotide molecule.
9. The method of claim 1 , further comprising identifying polynucleotide molecules at one or more genetic loci comprising a sequence variant.
10. The method of claim 1 , further comprising calculating a quantitative measure of paired sequence reads that map to a genetic locus, wherein both strands of said paired sequence reads comprise a sequence variant.
11. The method of claim 1 , further comprising calculating a quantitative measure of unpaired molecules bearing a sequence variant.
12. The method of claim 1 , wherein the tagging step further comprises attaching adapters to ends of each of said double-stranded DNA molecules in said biological sample, wherein said adapters tag a 5′ end of a strand of an individual double-stranded DNA molecule among said double-stranded DNA molecules with a first tag and a 3′ end of a complementary strand of said individual double-stranded DNA molecule with a second tag, thereby providing said tagged strands.
13. The method of claim 12 , further comprising (i) sequencing at least a portion of said tagged strands to produce a set of raw sequence reads, and (ii) mapping said set of raw sequence reads to a genetic locus in a reference genome, wherein said first tag and said second tag are indicative of which strand of said tagged strands each of said set of raw sequence reads is derived.
14. The method of claim 12 , wherein said adapters are from a set of library adapters comprising a plurality of polynucleotide molecules with said tag sequences.
15. The method of claim 1 , wherein the tag sequences are of between approximately 3 and 20 nucleotides in length.
16. The method of claim 12 , wherein said adapters do not include sequencing primer binding sites.
17. The method of claim 12 , wherein said adapters each include a fixed sequence at a ligatable end of the adapter.
18. The method of claim 1 , wherein the tagging step further comprises attaching adapters to ends of each of said double-stranded DNA molecules in said biological sample, and wherein said tag sequences are sufficiently unique so that individually tagged double-stranded DNA molecules can be differentiated from other tagged double-stranded DNA molecules.
19. The method of claim 1 , wherein the tagging step further comprises attaching adapters from an adapter set to ends of each of said double-stranded DNA molecules in said biological sample, wherein the adapter set includes adapters having non-unique tag sequences, and wherein said tagged double-stranded DNA molecules can be differentiated from other tagged double-stranded DNA molecules using—
(i) a combination of a non-unique tag sequence and said information from beginning and/or end portions of said raw sequences of said parent polynucleotides, or
(ii) a combination of a first non-unique tag sequence at a first end of said double-stranded DNA molecule and a second non-unique tag sequence at a second end of said double-stranded DNA molecule, or
(iii) a combination of (i) and (ii).
20. The method of claim 1 , further comprising, prior to (c), subjecting said double-stranded DNA molecules, said tagged strands, or said enriched tagged strands to amplification.
21. The method of claim 1 , further comprising calculating a copy number variation for a genetic locus in said set of one or more genetic loci, wherein said copy number variation is based at least in part on the quantification of said enriched tagged strands that map to said genetic locus for which complementary strands are detected and the quantification of said enriched tagged strands that map to said genetic locus for which only one strand among complementary strands is detected.