Bwa Alignment: Gap Vs Soft-Clipped Sequences

Question

[michealsmith]

Suppose we have a 30bp-50bp deletion, and read length is 100bp. If the read happens to cover the deletion, I'm just wondering, will BWA align the read with soft-clipped sequences, or allow this 30bp-50bp gap?

I'm pretty much sure for deletion like 10bp or even up to 20bp, BWA can do the gapped alignment; and for deletions, say 70, 80bp, BWA will produce split reads. But I'm just confused about this middle state.

Thanks a lot!

Answer 1

[Leonor Palmeira]

This is very dependent on the options you use on your 'bwa aln' incantation. Although I don't believe 'bwa' was designed to map reads with such large indels. From my experience, I use it for mappings with a handful of indels, not much more.

This being said, you can tweak some options to allow for more indels. You can for instance have a look at these two options on 'bwa aln':

• -o INT maximum number or fraction of gap opens
• -e INT maximum number of gap extensions

Otherwise, you can also use novoalign, stampy or blat which may handle large indels better than bwa.

[michealsmith]

thanks. But can you explain "fraction of gap opens" and "gap extensions" more?