minimap2 icon indicating copy to clipboard operation
minimap2 copied to clipboard

Mention in man page that CIGAR strings in SA tags are approximate

Open fedarko opened this issue 1 month ago • 0 comments

As documented in #724, the CIGAR strings used to describe supplemental alignments in the SA tags produced by minimap2 are approximate—for example, a CIGAR string of 23M 1I 537M 1I 64M 1I 512M 5180H for an alignment is turned into 1136M 3I 5180S when this alignment is represented within a SA tag.

This property of SA tags is currently undocumented, both in minimap2's code and in the SAM optional fields specification. As far as I can tell, multiple people have been confused by this over the years (#724, #524, #406, #287, https://github.com/eldariont/svim/issues/5).

To rectify this issue, this PR adds a short line to minimap2's man page mentioning that CIGAR strings are approximate. I'd also be happy to add an entry in the FAQ describing this issue, but I think just updating the man page (which is where the README says to check for "detailed description of minimap2 command line options and optional tags") should be sufficient for now.

This PR also fixes an unrelated small typo in the FAQ (impriciseimprecise).

fedarko avatar May 18 '24 04:05 fedarko