Personalized genome assembly for accurate cancer somatic mutation discovery using cancer-normal paired reference samples

Abstract The use of personalized genome assembly as a reference for detecting the full spectrum of somatic events from cancers has long been advocated but never been systematically investigated. Here we address the critical need of assessing the accuracy of somatic mutation detection using personalized genome assembly versus the standard human reference assembly (i.e. GRCh38). We first obtained massive whole genome sequencing data using multiple sequencing technologies, and then performed de novo assembly of the first tumor-normal paired genomes, both nuclear and mitochondrial, derived from the same donor with triple negative breast cancer. Compared to standard human reference assembly, the haplotype phased chromosomal-scale personalized genome was best demonstrated with individual specific haplotypes for some complex regions and medical relevant genes. We then used this well-assembled personalized genome as a reference for read mapping and somatic variant discovery. We showed that the personalized genome assembly results in better alignments of sequencing reads and more accurate somatic mutation calls. Direct comparison of mitochondrial genomes led to discovery of unreported nonsynonymous somatic mutations. Our findings provided a unique resource and proved the necessity of personalized genome assembly as a reference in improving somatic mutation detection at personal genome level not only for breast cancer reference samples, but also potentially for other cancers.