Abstract
Low-frequency somatic mutations accumulate in normal tissues throughout life and contribute to cancer initiation, yet their detection is limited by the sensitivity of conventional sequencing methods. We describe CarcSeq, an error-corrected, targeted next-generation sequencing platform developed to quantify rare cancer driver mutations (CDMs) at variant allele frequencies as low as ~10-4. CarcSeq integrates high-fidelity amplification, unique molecular identifiers, and single-strand consensus sequencing to accurately measure mutant frequency (MF) across a curated panel of sequences encompassing recurrent oncogene and tumor suppressor hotspots. To capture mutation heterogeneity associated with early clonal expansion, the median absolute deviation (MAD) of MF was incorporated as a metric.
Application of CarcSeq to normal and tumor-adjacent human tissues revealed tissue-specific mutational profiles and demonstrated that recurrent driver mutations are detectable in histologically normal samples. Extension to rodent models showed that MAD correlates strongly with strain-, tissue-, and sex-specific spontaneous tumor incidence, supporting its utility as an early biomarker of neoplastic susceptibility. In an exposure study, CarcSeq detected dose- and time-dependent clonal expansion of spontaneous Pik3ca H1047R mutations following administration of the nongenotoxic carcinogen lorcaserin, despite no overall increase in global mutation frequency, highlighting sensitivity to early carcinogenic processes not captured by traditional genotoxicity assays.
Compared with whole-genome, whole-exome, and ultra-deep error-corrected sequencing approaches, CarcSeq balances sensitivity, throughput, and cost by focusing on biologically human cancer-relevant driver mutation hotspots. Together, these findings establish CarcSeq-derived MF and MAD as quantitative, cross-species biomarkers of early clonal expansion with applications in translational carcinogenicity assessment, drug development, and cancer risk modeling.
Keywords
Next-generation sequencing, Cancer, Mutations, CarcSeq, Tumor