PT  - JOURNAL ARTICLE
AU  - Alexander T Dilthey
AU  - Pierre-Antoine Gourraud
AU  - Zamin Iqbal
AU  - Gil McVean
TI  - High-accuracy HLA type inference from whole-genome sequencing data
AID  - 10.1101/035253
DP  - 2015 Jan 01
TA  - bioRxiv
PG  - 035253
4099  - http://biorxiv.org/content/early/2015/12/24/035253.short
4100  - http://biorxiv.org/content/early/2015/12/24/035253.full
AB  - Extensive hyperpolymorphism and sequence similarity between the HLA genes make HLA type inference from whole-genome sequencing data a challenging problem. We address these by representing sequences from over 10,000 known alleles in a reference graph structure, enabling accurate read mapping. HLA*PRG, our algorithm, outperforms existing methods by a wide margin and for the first time consistently achieves the accuracy of gold-standard reference methods with one error across 158 alleles tested.