pysamstats counting splicing events as deletions

[saldi]

Hello!


I am using pysamstats to calculate the number of mismatches, deletions and insertions per base in my RNAseq data. The RNA was treated with a chemical such that sites which are single stranded end up getting mutated, deleted etc. However, I am running into a problem as pysamstats is counting spliced regions as deletions. Does anyone know of a way I could fix this problem? Maybe a way I could get pysamstats to only consider deletions smaller than 4bp? Or some other program that could give me this information in a tab format?

Thanks!

[Brian Bushnell]

Do these sites appear in your sam file cigar strings as an "N" or a "D"?

[saldi]

They appear to be N. The reads with "real" deletions have D. IGV is counting them correctly (if I hoover my cursor over that position in the display, I get the correct number of deletions based on manually counting them in the displayed reads), but pysamstats gives all the deletions plus the number of splicing events that cross that position.

Maybe if there was a way to count the number of splicing events crossing each position I could match that to the pysamstats output and use awk to subtract out the splicing events from the deletion total?

Thanks!

[Brian Bushnell]

It's worth trying... I don't have a better suggestion.

[saldi]

OK- do you happen to know how I could count number of splicing events that cross each position?

Thanks!

[Brian Bushnell]

Well, no, unfortunately...

But BBMap's reformat.sh does support this flag:

dellenfilter=-1         Discard reads with a deletion longer than this.

Perhaps that would be useful? It considers both N and D as deletion events, but assuming that you can separate them based on length, that should work, for most mappers. Then you can call the short deletions on the output with a variant caller and ignore them subsequently.