Haplotype-aware segmentation with HapASeg increases accuracy of detecting homolog-specific somatic copy number alterations

Abstract Somatic copy number alterations (sCNAs) drive cancer initiation, progression, resistance, and metastasis. Furthering our understanding of sCNAs requires substantially larger cohorts. Most tumors available for sequencing are preserved with formalin-fixed, paraffin-embedding (FFPE), which causes DNA cross-linking that distorts coverage profiles and challenges current sCNA estimation methods. Traditional methods denoise data using large panels of similar normal samples, which are impractical to obtain for FFPE cohorts. Here, HapASeg overcomes this limitation by leveraging haplotype phasing and unique covariates to accurately estimate sCNA segments across FFPE, fresh frozen, whole genome sequencing and whole exome sequencing sample types, outperforming current methods without requiring panel-of-normal correction.