Single molecule molecular inversion probes for targeted, high-accuracy detection of low-frequency variation

Abstract

The detection and quantification of genetic heterogeneity in populations of cells is fundamentally important to diverse fields, ranging from microbial evolution to human cancer genetics. However, despite the cost and throughput advances associated with massively parallel sequencing, it remains challenging to reliably detect mutations that are present at a low relative abundance in a given DNA sample. Here we describe smMIP, an assay that combines single molecule tagging with multiplex targeted capture to enable practical and highly sensitive detection of low-frequency or subclonal variation. To demonstrate the potential of the method, we simultaneously resequenced 33 clinically informative cancer genes in eight cell line and 45 clinical cancer samples. Single molecule tagging facilitated extremely accurate consensus calling, with an estimated per-base error rate of 8.4 × 10−6 in cell lines and 2.6 × 10−5 in clinical specimens. False-positive mutations in the single molecule consensus base-calls exhibited patterns predominantly consistent with DNA damage, including 8-oxo-guanine and spontaneous deamination of cytosine. Based on mixing experiments with cell line samples, sensitivity for mutations above 1% frequency was 83% with no false positives. At clinically informative sites, we identified seven low-frequency point mutations (0.2%–4.7%), including BRAF p.V600E (melanoma, 0.2% alternate allele frequency), KRAS p.G12V (lung, 0.6%), JAK2 p.V617F (melanoma, colon, two lung, 0.3%–1.4%), and NRAS p.Q61R (colon, 4.7%). We anticipate that smMIP will be broadly adoptable as a practical and effective method for accurately detecting low-frequency mutations in both research and clinical settings.

MIPSTR: a method for multiplex genotyping of germline and somatic STR variation across many individuals

“…1A; Hiatt et al 2013). In Col-0, we successfully captured all 102 STR target loci (Supplemental Fig.…”

“…We designed MIPs to capture 200 bp, a larger target size than is typically used (Porreca et al 2007;Turner et al 2009;O'Roak et al 2012;Hiatt et al 2013), to increase the size range of targeted STRs. This size range encompasses the majority of STRs (Gymrek et al 2012), except those that have undergone extreme expansion as seen in some human diseases and for the intronic STR in the A. thaliana IIL1 gene (Gatchel and Zoghbi 2005;Sureshkumar et al 2009).…”

“…We used single-molecule Molecular Inversion Probes (smMIPs) (Hiatt et al 2013) to capture STRs, thereby maximizing the number of STR-spanning, informative reads. In a proof-of-principle experiment, we targeted 102 STRs across the genome of the model (Sullivan et al 2014), and intergenic tri-and hexa-nucleotide STRs (Supplemental Fig.…”

“…MIPSTR combines STR capture via single-molecule Molecular Inversion Probes (smMIPs) (Hiatt et al 2013) with midsize sequencing reads (250 bp) and a unique mapping strategy. In proof-of-principle experiments, we captured and sequenced STRs genome-wide in diverse A. thaliana populations, called germline STR genotypes with high accuracy, and quantified technical error with singlemolecule information.…”

Facilitated sequence counting and assembly by template mutagenesis

“…There are several protocols in which a sequence of random nucleotides is appended to the template molecules before amplification and sequencing. This idea has been applied under a variety of names to identify PCR duplicates (1,2), improve counting of DNA (3,4) and RNA (5-7) templates, and reduce sequence error (8)(9)(10). Each implementation has its own name for the random nucleotide sequences, and we refer to them as varietal tags (11).…”