HLA

xHLA

I've run this on the end-to-end output aligning to hg38 reference containing alternates with bowtie2 and compared it to the output of aligning to just hg38 chr6 with bwa and there was only 1 difference for the entire data set.

Note that for some reason one of the R scripts crashes on cluster as demands too much memory, but I've never seen it use more than 30GB?