%0 Journal Article %A Lin Huang %A Petr Danecek %A Sivan Bercovici %A Serafim Batzoglou %T Large-scale Population Genotyping from Low-coverage Sequencing Data using a Reference Panel %D 2016 %R 10.1101/085936 %J bioRxiv %P 085936 %X In recent years, several large-scale whole-genome sequencing projects were completed, including the 1000 Genomes Project, and the UK10K Cohorts Project. These projects aim to provide high-quality genotypes for a large number of whole genomes in a cost-efficient manner, by sequencing each genome at low coverage and subsequently identifying alleles jointly in the entire cohort. The resulting variant data are critical for the characterization of human genome variation within and across populations in the original projects, and for many downstream applications such as genome-wide association studies. The same datasets carry the potential to increase the quality of genotype calling in other low-coverage sequencing data in future sequencing projects, because the existing genotype calls capture the linkage disequilibrium structures of the cohorts they represent. In this paper we present reference-based Reveel (or Ref-Reveel in short), a novel method for large-scale population genotyping. Ref-Reveel is based on our earlier method, Reveel, which has been demonstrated to be an effective tool for variant calling from low-coverage sequencing data. Ref-Reveel leverages genotype calls from a sequenced cohort to boost the genotyping quality of new datasets, while maintaining high computational efficiency. We show that using a reference panel improves the quality of genotype calling via extensive experiments on simulated as well as real whole-genome data. Ref-Reveel is publicly and freely available at http://reveel.stanford.edu. %U https://www.biorxiv.org/content/biorxiv/early/2016/11/05/085936.full.pdf