I would like to know how BWA mem handles repetitive hits.
I know that bwa aln/samse- given a read that maps to multiple regions- will randomly select one region. Each query will be listed once in the sam file, and alternative mapping locations for that read will be listed in XA if the number of hits is less than INT hits.
But what about BWA mem? I keep reading that BWA mem differs from bwa aln in that it will split queries, but what about reads that aren't split but still map to multiple locations? Does mem also chose a random location? Are multiply mapped reads reported in the same fashion in bwa mem as they are for bwa aln/samse (that is, with one listing per query and XAs reported if less than INT)?
For simplicity's sake, I am not interested in using other aligners. I have pre-trimmed RAD data (80-139bp) from a frog species with a highly repetitive genome.