If you’re interested in immunogenetics, you probably know that high-accuracy HLA type inference from all types of whole-genome sequencing data is now possible: we recently published a tool, HLA*PRG, that enables that.

You might also have noticed, however, that HLA*PRG can be quite resource-hungry – hundreds of CPU hours and up to 80GB of RAM for 2 x 250bp whole-genome sequencing data. Also, it only supported GRCh37.

In this post I describe an algorithm, HLA*PRG:LA, that brings down these requirements to ~30GB and ~30 CPU hours per sample, without sacrificing accuracy.

Continue reading on the Genome Informatics blog…