Weird alignments in amplicon sequencing

Question

[kulvait]

Hi, we are using Illumina Myeloid panel of ~500 amplicons to sequence certain part of exome by Illumina HiSeq. I were doing alignment using bwa mem with default parameters. In some regions the quality of mapping is very poor. In the first case there is lot of mismatches and in the second case there is lot of soft clipping or mismatches at the ends of some reads. Is it possible to somehow force bwa mem to discard these mappings and is it wise? I mean would you set some speciffic parameters to bwa mem mapping to avoid these or are these virtually unavoidable?

Thanks Vojtech.

[John]

Hi, what aligner do you use? You could try to increase in bwa-mem parameter B - penalty for mismatch and -L penalty for 5' and 3' end clipping.

[kulvait]

I have been using bwa mem with default parameters. This problem affects minority of the data. Thank you for recomendation, do you have experience with particular setting, we will consider using that.