Calculate alignment similarity

[cmorganl]

It would be nice to have the option of filtering read alignments by percent similarity, and potentially even reporting summary values if a user desired. For example, a separate output table with the mean and/or median percent similarity of the reads used to calculate the abundance metrics for a contig.

Heng Li described three different alignment similarity equations on his blog. I am personally leaning towards the one he favours, gap-compressed identity, over the BLAST and gapless versions.

The similarity between a read and the mapped reference sequence can be calculated directly from the CIGAR string and should be done in the C++ extension to keep this as quick as possible.

[RyloByte]

I agree with you here, this would be a really great feature!