mismatch frequency

[MARIAC]

Hello,
I'm looking for a script to extract from sam file the number of M, (match base in the CIGAR string) and the number mismatch (NM:i field).
The objective is to calculate a mismatch frequency in order to filter reads.

Thanks

Cheers
Cédric

[dpryan]

Here's a very simple python script using pysam that demonstrates how to do this.

[amitm]

hi,
The "M" in CIGAR doesn't necessarily mean match; it means alignment. That alignment can be a match or mismatch.
The NM:i field would contain valid mismatches but also insertions and deletions.

Check -

http://samtools.github.io/hts-specs/SAMv1.pdf

BWA has a flag to restrict no. of edit pos. (mismatch +InDels). So if considering InDels too suits you then you can use the flag and filtering is done automatically.