GVC: A superfast and universal genomic variant caller

Abstract

Germline and somatic variant detection from human and cancer whole-genome sequencing data is a challenge task for genome-wide association study and cancer genomics in precision medicine. Many confounding factors contribute the difficulties including complexity of variant, sequencing and alignment error, tumor clonality and sample purity etc. Current genomic variant callers are too time-consuming to meet the requirement of clinical application in precision medicine. We developed superfast and universal Genomic Variant Caller (GVC), which can simultaneously detect various genomic variants including SNV, sINDEL and SV from personal and normal-cancer paired whole-genome/exome sequencing data within fifteen minutes. Whats more, it achieved higher sensitivity and precision than popular variant callers including GATK4, Mutect, NovoBreak in germline and somatic variant detection from NA12878 and ICGC-TCGA Dream Challenge Datasets respectuvely. It is worth mentioning that GVC achieved comparable performance in variant detection from NA12878 sequenced by three different high-throughput sequencing platforms including Illumina HiSeq2000, NovaSeq and BGISEQ-500.