Sunday, 17 March 2013

A call for params used for bwa mem for variant calling applications



---------- Forwarded message ----------
From: Sean Davis
Date: Sunday, 17 March 2013
Subject: [Bio-bwa-help] bwa mem for variant calling applications
To: BWA Help List <bio-bwa-helps.sourceforge.net>


Hi, all.

We are playing with bwa mem as a potential replacement for bwa aln in variant calling applications for 70-150bp pe reads because of increased speed, increased sensitivity, and for the potential for being more sensitive in finding some unusual variants, particularly 10-50bp indels.  However, we are finding that the default parameters result in soft clipping to an extent that many true variants are simply soft-clipped out.  I know other folks are probably experimenting the same way and wondered what tuning people are trying.

I'll say that I personally hate "what parameters should I use questions", but I thought at this early stage in bwa mem experimentation, we could benefit from each others' experiences.  

Thanks,
Sean


No comments:

Post a Comment

Datanami, Woe be me