Gap penalty aligner
Stephan Lorenzen
SNAP (1.0beta18) announces -G for single end reads (specify a gap penalty to use when generating CIGAR strings) in the help; however in a test I saw no effect and in the source code only a setting of a variable in AlignerOptions.cpp. Is this feature planned for the future? Will it have any effect on the alignment chosen (e.g. in case of equal edit distance) of just on the CIGAR string generated for some given match?
Thanks,
Stephan
Bill Bolosky
Yes, I’m working on it. I’d forgotten that I’d let that get into the main branch, it must have tagged along with some other changes. As far as I can tell, SNAP’s main source of inaccuracy is that it’s incorrectly calling short indels near the ends of reads. It only does this if that’s the only way to get the smallest possible edit distance, but sometimes it’s still not what you want, hence the affine gap edit distance computations.
The idea is to allow it to be used both for CIGAR string and also for alignment depending on other command line options. I worry a little that using it in alignments might have a noticeable performance penalty, since that would go right in the heat of the main loop.
The bad news is that I haven’t had a lot of time to work on it lately, so I’m not sure when it will be ready and I wasn’t planning on making any announcements about it until you found that switch.
--
You received this message because you are subscribed to the Google Groups "SNAP Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
snap-user+...@googlegroups.com.
To post to this group, send email to
snap...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/snap-user/a97471d5-64c6-441d-88ff-9c3c4152d63a%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.