search

samtools / hts-specs

Specifications of SAM/BAM and related high-throughput sequencing file formats
http://samtools.github.io/hts-specs/
655 stars 172 forks source link

Meaning of "next segment" in FLAG #81

Open nathanhaigh opened 9 years ago

nathanhaigh commented 9 years ago

For the RNEXT and PNEXT fileds the spec refers to the "mate/next read". Further, for RNEXT it states:

For the last read, the next read is the first read in the template.

Thus creating a circular linked reference for multiple reads/segments from the same template. Is the same meaning of "next" to also be taken for the sam flags? For example, should 0x8 (MUNMAP) and 0x20 (MREVERSE) of the last read be in reference to the first read in the template?

This has implications for tools calculating the orientation of read pairs.

Cheers, Nathan

jkbonfield commented 9 years ago

I can't say anything on the wording of the spec without looking at it for longer, but IMO the meaning of flag 0x20/0x8 should be coupled to the meaning of pnext/rnext. It is pointless to get into situations where they refer to different reads as it makes both of them less useful. Unfortunately one indicates primary only reads while the other does not. I suspect this is just an omission rather than a deliberate policy.

As far as the general policy goes of the flags working in a circular linked list, this is true. It's just some nuances which right now are poorly defined.