bwa aln for long gaps

[ehlin]

Hello everyone, and thank you in advance for your help.

I am trying to align Illumina Ion Torrent reads to a reference genome, and these reads are in the TCR region. This means that there may be long gaps in between reads due to VDJ recombination (such as 1 read that is supposed to be from 7:1-100 and then 7:800-900, etc.) I wanted bwa aln to account for this by adjusting gap extension penalty to 0, and allowing 600,000 gap extensions to cover the whole region. It doesn't seem to be working, does anyone have any insights?

[swbarnes2]

I don't think that bwa is going to map reads like that.

It would have to be splitting the read up and doing two alignments, and it doesn't do that. The longest indel I've seen it was was about 10 bases.

bwa would tell you that you would have no coverage in the gap, and it might align some reads only by soft clipping the overhang, so you could figure it out from that, given enough time, but bwa is never going to give you an alignment with a 800 bp gap.

[ehlin]

I don't think that bwa is going to map reads like that.

It would have to be splitting the read up and doing two alignments, and it doesn't do that. The longest indel I've seen it was was about 10 bases.

bwa would tell you that you would have no coverage in the gap, and it might align some reads only by soft clipping the overhang, so you could figure it out from that, given enough time, but bwa is never going to give you an alignment with a 800 bp gap.

I see, that is unfortunate... but better that I know, since I tried a few more times with SAM files with 0 bytes, I'll keep that in mind. Thank you so much for your help!

-Edwin