Low Frequency Variant Detection In Cell Free Dna By Applying Molecular Identifiers To Targeted Ngs

Abstract The growing use of liquid biopsy for early detection and monitoring of disease necessitates accurate variant detection at <1% allele frequencies due to a low population of disease DNA within circulating, cell-free DNA (cfDNA). Reliable, low-frequency variant detection by next-generation sequencing (NGS) is challenging due to background noise from PCR and sequencing errors. We employed molecular identifiers (MIDs) to uniquely label individual DNA molecules prior to amplification, facilitating the distinction of true variants from PCR and sequencing errors. We incorporated MIDs in both our amplicon library prep that uses multiplex PCR for targeted NGS and our whole genome library prep followed by targeting with hybridization capture using an 800kb pan-cancer panel. We performed low frequency spike-in experiments at <1% allele frequencies. We prepared MID libraries with various amplicon panels including a 17 amplicon EGFR pathway panel and a 104 amplicon SNP panel. Deep sequencing to >30,000x was done to maximize MID family size (number of PCR duplicates) and optimize generation of a consensus sequence. This analysis identified all known variants present at 1%, 0.5%, and 0.25% allele frequencies. Next, the hybridization capture libraries were prepared with low-frequency spike-in samples, sequenced to >8000x, and all known variants at 1% and 0.5% allele frequencies were maintained in the consensus data. In both cases, the number of false positives was reduced, resulting in improved specificity. Further, EGFR amplicon libraries and hybridization capture libraries were prepared using cfDNA samples from lung, ovarian, liver, stomach, and colon cancers. Variant calling based on MID generated consensus sequences identified mutations in cfDNA samples as well as corresponding tumor and normal samples when available. This study highlights the ability of MID technology to enable low frequency variant detection, critical to track known variants and identify novel pathogenic mutations in cfDNA samples. Citation Format: Ashley Wood, Sukhinder Sandhu, Mida Pezeshkian, Vanessa Kelchner, Jordan RoseFigura, Justin Lenhart, Laurie Kurihara, Vladimir Makarov. Low frequency variant detection in cell free DNA by applying molecular identifiers to targeted NGS [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2230.