No soft-clipping in BWA 0.7.10 anymore?

Question

[Sebastian]

Hi all!

After moving from bwa version 0.5.8c to 0.7.10 I discovered differences that look like errors in the newer version: it seems as BWA doesn't perform soft-clipping anymore, resulting in false positive variant calls:

Is that true? I could find parameters to disable soft-clipping, but no parameter to explicitly turn on soft-clipping.

Both version were used with default parameters (both, bwa aln and bwa sampe).

Any help would be appreciated!

Thank you in advance,
Sebastian

[Devon Ryan]

Hello Sebastian!

It appears that your post has been cross-posted to another site: BWA mailing list and SEQanswers.

This is typically not recommended as it runs the risk of annoying people in both communities.

[Devon Ryan]

Give Heng Li a chance to reply to your email from yesterday.

[Sebastian]

Yes you are right, I just wanted to reach some more users...

Answer 1

[Devon Ryan]

I just tried version 0.7.12 via bwa-meth with a test dataset that requires soft-clipping from bison and got a lot of soft-clipped alignments...as expected. So at the very least the most recent version works.

Answer 2

[Peter]

Duplicates are

• http://seqanswers.com/forums/showthread.php?t=50050 and
• http://sourceforge.net/p/bio-bwa/mailman/message/33329232/ where Heng Li replied it was a bug fixed in 0.7.12

[Sebastian]

It seems to be another bug, see thread on seqanswers (sorry for cross-posting!)...