Identification of somatic or germline origin for cell-free DNA
The present disclosure provides systems and methods to detect somatic or germline variants from cell-free DNA (cfDNA). Generally, the systems and methods comprise receiving sequencing information from cfDNA from said subject, determining whether measures are above or below a threshold; and classifying each as being of somatic origin or classifying each locus as being of germline origin.
1 . A method of identifying a germline origin or a somatic origin of a genetic variant comprising:
allocating a read budget of sequence reads that specifies a first number of sequence reads directed to a backbone sub-panel that interrogates a plurality of backbone genomic loci and a second number of sequencing reads directed to a hotspot sub-panel that interrogates a plurality of hotspot genomic loci, wherein the second number is selected to achieve a user-defined lower-limit-of-detection;
a) receiving, from a sequencer, sequence data indicative of a plurality of sequence reads of cfDNA molecules generated in accordance with the allocated read budget, wherein the sequence data comprises sequence data for the plurality of backbone genomic loci and the plurality of hotspot genomic loci;
b) processing, using a computer, the sequence data to:
(i) for each of the plurality of backbone genomic loci, obtain an allele fraction (AF) measure and a measure of AF variability,
(ii) bin the plurality of backbone genomic loci based on a presumed origin, and
(iii) using the AF measures and the measures of AF variability of the binned plurality of backbone genomic loci, determine an AF variability threshold such that backbone genomic loci of the binned plurality of backbone genomic loci having AF variability above the AF variability threshold are designated high copy number loci and backbone genomic loci of the binned plurality of backbone genomic loci having AF variability below the AF variability threshold are designated low copy number loci and an AF threshold that separates germline from somatic AF distributions associated with the backbone genomic loci of the binned plurality of backbone genomic loci designated as the low copy number loci;
c) for a variant detected at any hotspot genomic locus of the plurality of hotspot genomic loci, obtaining, using a computer, a variant AF and a variant AF variability, and comparing the variant AF variability against the AF variability threshold to identify the variant with the measure of variant AF variability below the AF variability threshold as having low copy number; and
d) based on the variant having a low copy number, identifying germline origin or somatic origin for the variant by comparing, using a computer, the variant AF against the AF threshold, wherein,
the variant is germline origin if the variant AF is above the AF threshold, and
the variant is somatic origin if the variant AF is below the AF threshold.
2 . The method of claim 1 , further comprising identifying a copy number variation (CNV) level, wherein said measure of AF variability for a genomic locus below said AF variability threshold indicates a low CNV for said genomic locus, and wherein said measure of AF variability for a genomic locus above said AF variability threshold indicates a high CNV for said genomic locus.
3 . The method of claim 1 , wherein the AF threshold is derived from a second AF measure of a variant of a known origin, wherein said known origin is (i) somatic or (ii) germline.
4 . The method of claim 3 , wherein said second AF measure is derived from said sequence data for a second genomic locus from said plurality of backbone genomic loci.
5 . The method of claim 3 , wherein said second AF measure is derived from a genomic locus from cfDNA from a control subject.
6 . The method of claim 3 , wherein d) comprises identifying said variant as being of the same origin as said known origin, if there is a difference of 10% or less between the variant AF and the AF threshold.
7 . The method of claim 1 , further comprising assigning a presumed origin of each locus of said plurality of backbone genomic loci, wherein said presumed origin is one of: (i) presumed somatic origin, (ii) presumed germline origin, and (iii) presumed indeterminate origin.
8 . The method of claim 7 , wherein binning said plurality of backbone genomic loci based on said presumed origin, generates a first bin comprising genomic loci of presumed somatic origin, a second bin comprising genomic loci of presumed germline origin, and a third bin comprising genomic loci of indeterminate origin;
and determining, for one or more genomic regions in each of the first bin, the second bin, and the third bin, a quantitative AF measure for one or more loci in each of the one or more genomic regions based on cfDNA sequencing reads to generate a first AF set, a second AF set, and a third AF set, respectively.
9 . The method of claim 8 , wherein the AF threshold is determined by:
(a) generating a first frequency distribution based on said first AF set and a second frequency distribution based on said second AF set, wherein no overlap exists between said first frequency distribution and said second frequency distribution;
(b) identifying the AF threshold based on said first and second frequency distributions, which AF threshold is (i) no less than the largest quantitative AF measure among said first AF set and (ii) no more than the smallest quantitative AF measure among said second AF set.
10 . The method of claim 8 , wherein the AF threshold is determined by (i) generating a first frequency distribution based on said first AF set and a second frequency distribution based on said second AF set, wherein an overlap exists between said first frequency distribution and said second frequency distribution and (ii) identifying the AF threshold based on said first and second frequency distributions, which the AF threshold is the largest quantitative AF measure among said first AF set;
wherein the method further comprises identifying a second AF threshold based on said first and second frequency distributions, which second AF threshold is the smallest quantitative AF measure among said second AF set; and
wherein d) comprises identifying each backbone genomic locus of the plurality of backbone genomic loci as:
(A) somatic origin if said AF measure is no more than the AF threshold,
(B) germline origin if said AF measure is no less than said second AF threshold, or
(C) ambiguous if said AF measure is greater than the AF threshold and less than the second AF threshold.
11 . The method of claim 1 , wherein said measure of variability is a standard deviation of said AF measure.
12 . The method of claim 1 , wherein a genomic locus is selected from a member of the group consisting of BRAF, BRCA1, BRCA2, EGFR, KRAS, PIK3CA, ROS1, and TP53.
13 . The method of claim 1 , further comprising prior to a), providing a plurality of cfDNA molecules.
14 . The method of claim 13 , further comprising amplifying said plurality of cfDNA molecules.
15 . The method of claim 13 , further comprising enriching for a subset of said plurality of cfDNA molecules, wherein said subset comprises cfDNA molecules comprising a sequence of a genomic region of interest.
16 . The method of claim 13 , further comprising tagging said plurality of cfDNA molecules.
17 . The method of claim 13 , further comprising sequencing said cfDNA molecules to generate said sequence reads of cfDNA.
18 . The method of claim 1 , further comprising aligning said set of sequence reads to a genetic locus.
19 . The method of claim 1 , further comprising mapping said sequence reads to a reference genome.
20 . The method of claim 1 , wherein binning comprises use of a computer database.
21 . The method of claim 1 , comprising obtaining a sample and generating a set of at least 1 million sequence reads.
22 . The method of claim 1 , comprising a read depth of about 5,000×.
23 . The method of claim 1 , comprising a read depth of about 10,000×.
24 . The method of claim 1 , comprising a read depth of about 20,000×.
25 . The method of claim 1 , wherein a targeted-sequencing panel comprises the backbone sub-panel and the hotspot sub-panel and encompasses a plurality of genes or genomic regions such that a predetermined proportion of cancer patients exhibits a genetic variant in at least one region of the targeted-sequencing panel.
26 . The method of claim 25 , wherein capture probes for the targeted-sequencing panel comprise (i) probes directed to hotspot regions and (ii) nucleosome-aware probes that optimize capture efficiency based on cfDNA fragment-size variation and GC-content.
27 . The method of claim 25 , wherein the targeted-sequencing panel comprises a plurality of sub-panels comprising one or more of: (a) a tissue-of-origin sub-panel, (b) a whole-genome scaffold sub-panel, and (c) a transcription-start-site/CpG-island sub-panel.
28 . The method of claim 1 , wherein the read budget is between 100 million and 100 billion sequence reads.
29 . The method of claim 1 , wherein the plurality of backbone genomic loci are sequenced at a first average read depth and the hotspot loci are sequenced at a second, deeper average read depth.
30 . The method of claim 1 , wherein nucleic acids from two or more subjects are pooled and sequenced in a single run at ratios determined by amounts of nucleic acid obtained from each subject.