Detecting mutations for cancer screening

A method for identifying somatic mutations in a human subject by analyzing a biological sample of the human subject, the biological sample including DNA fragments originating from normal cells and potentially from tumor cells or cells associated with cancer, the biological sample including cell-free DNA fragments, the method comprising:

• obtaining template cell-free DNA fragments from the biological sample to be analyzed;
• preparing a sequencing library of analyzable cell-free DNA molecules using the template cell-free DNA fragments, wherein a duplication rate of the sequencing library from the template cell-free DNA fragments is less than 5%;
• sequencing the sequencing library of analyzable cell-free DNA molecules to obtain a plurality of sequence reads of cell-free DNA molecules;
• receiving, at a computer system, the plurality of sequence reads;
• aligning, by the computer system, the plurality of sequence reads to a reference human genome to determine genomic positions for the plurality of sequence reads;
• obtaining, by the computer system, information about a constitutional genome corresponding to the human subject;
• and
• comparing, by the computer system, the plurality of sequence reads of cell-free DNA molecules to the constitutional genome to identify a filtered set of loci as having somatic mutations in some tissue of the human subject, wherein:
• at each locus of the filtered set, a number of the plurality of sequence reads of cell-free DNA molecules having a sequence variant relative to the constitutional genome is above a cutoff value, the cutoff value being greater than one.

A method for identifying somatic mutations in a human subject by analyzing a biological sample of the human subject, the biological sample including DNA fragments originating from normal cells and potentially from tumor cells or cells associated with cancer, the biological sample including cell-free DNA fragments, the method comprising, performing, by a computer system:

• obtaining information about a constitutional genome corresponding to the human subject;
• receiving one or more sequence reads for each of a plurality of cell-free DNA fragments in the biological sample, thereby obtaining a plurality of sequence reads of cell-free DNA fragments;
• aligning the plurality of sequence reads to a reference human genome using a first alignment procedure to determine genomic positions for the plurality of sequence reads;
• comparing the plurality of sequence reads of cell-free DNA fragments to the constitutional genome as part of identifying a filtered set of loci as having somatic mutations in some tissue of the human subject, wherein:
• at each locus of the filtered set, a number of the plurality of sequence reads of cell-free DNA molecules having a sequence variant relative to the constitutional genome is above a cutoff value, the cutoff value being greater than one;
• for each candidate locus of a set of candidate loci identified as potentially having a somatic mutation:
• for each sequence read of the sequence reads aligning to the candidate locus using the first alignment procedure and having the sequence variant:
• determining whether the sequence read aligns to the candidate locus using a second alignment procedure that uses a different matching algorithm than that used for the first alignment procedure;
• when the sequence read realigns to the candidate locus using the second alignment procedure, determining a mapping quality of the realignment for the second alignment procedure;
• comparing the mapping quality to a quality threshold;
• and
• determining whether to discard the sequence read based on the comparing of the mapping quality to the quality threshold, wherein the mapping quality being less than the quality threshold provides a higher likelihood of discarding the sequence read than the mapping quality being greater than the quality threshold, thereby obtaining a number of remaining sequence reads;
• comparing the number of remaining sequence reads to a candidate threshold;
• and
• determining whether to discard the candidate locus based on the comparing of the number of remaining sequence reads to the candidate threshold, wherein the number of remaining sequence reads being less than the candidate threshold provides a higher likelihood of discarding the candidate locus than the number of remaining sequence reads being greater than the candidate threshold;
• and
• further identifying the filtered set of loci as having somatic mutations using the remaining candidate loci.