Reveel: large-scale population genotyping using low-coverage sequencing data

• DOI: 10.1093/bioinformatics/btv530
• Publication Date: 2015-09-10T13:37:06Z
• Authors: Lin Huang; Bo Wang; Ruitang Chen; Sivan Bercovici; Serafim Batzoglou
• Keywords: 1000 Genomes Project; Exome; Linkage Disequilibrium

Population low-coverage whole-genome sequencing is rapidly emerging as a prominent approach for discovering genomic variation and genotyping cohort. This combines substantially lower cost than full-coverage with discovery of low-allele frequency variants, to an extent that not possible array or exome sequencing. However, challenging computational problem arises jointly variants the entire Variant are relatively straightforward tasks on single individual has been sequenced at high coverage, because inference decomposes into independent each position which sufficient number confidently mapped reads available. in population sequencing, joint requires leveraging complex linkage disequilibrium (LD) patterns cohort compensate sparse missing data individual. The potentially massive computation time such inference, well confound low-frequency allele discovery, need be overcome this become practical.Here, we present Reveel, novel method nucleotide variant calling large cohorts have low coverage. Reveel introduces technique LD deviates from previous Markov-based models, aimed efficiency accuracy capturing rare haplotypes. We evaluate Reveel's performance through extensive simulations real 1000 Genomes Project, show it achieves higher state-of-the-art methods.http://reveel.stanford.edu/: [e-mail] available Bioinformatics online.