[FEATURE REQUEST] Add input-fmt-option to show number of match bases?

[y9c]

The document is a little confusion for qlen and rlen.

• qlen: no. query bases
• rlen: no. reference bases

It took me some time to realize that qlen includes match bases and soft-clipped bases, while rlen includes the intron skip. So, there is no parameters recording the number of matched bases, exclude the soft-clipped and intron-skip region. It is possible to have another parameter for "number of matched bases", say mlen?

Thanks!

[jkbonfield]

It doesn't sound too hard to add an extra parameter which is based on the CIGAR seq "match" count, but that would be the number of bases with a known alignment coordinate and not strictly matching as it's not possible to tell SNP from REF match using CIGAR alone. Note this would also treat "N" (ref skip) and "D" (deletion) equivalent too, as neither consume sequence bases.

Is that sufficient?

However note that's actually how it already works. It just calls the htslib bam_cigar2qlen and bam_cigar2rlen which is already doing everything you ask for except for also including soft-clips. So potentially it may be better resolved by a function to return the number of soft-clipped bases. That would then allow e.g. qlen - sclen.

Would that be preferable to a new term such as mlen which could be ambiguous: see cigar = vs cigar M opcode.

[y9c]

Thank you @jkbonfield for the suggestion. If a sclen function is available, that can also solve this isssue.