AS:i and XS:i SAM File Fields (Bowtie2)

[mapostolides]

A good explanation of those fields can be found here:

http://blog.kokocinski.net/index.php/sam-format-summary?blog=2

Basically, AS is the alignment score, and XS is the "suboptimal alignment score", essentially a secondary alignment. If AS and XS are close or equal, then you are getting multiple alignments happening.

[Brian Bushnell]

I suggest using required fields for analyses. It's not clear what the AS tag is supposed to report, and the XS tag is custom with no official documentation whatsoever. However, the required MAPQ field is clearly defined; it is the aligner's estimate that the alignment is correct. So, a MAPQ of 3 or lower means the aligner thinks there is a 50% chance or less that the alignment is correct. Uniquely-mapping reads, therefore, will have a MAPQ of 4 or higher.

[JuliaEpps]

I've seen posts/advice online stating that multimapper reads can be identified when the AS:i: and XS:i: optional fields in a .SAM file (after alignment with Bowtie2) are equal to one another such as AS:i:5 and XS:i:5 as opposed to the presence of XS:i: alone (which can be added to non-multimapper reads in some instances).

What are these fields actually describing? The manual for Bowtie2 is quite vague and, for example, simply describes AS:i: as the 'alignment score' - but from where is it derived? What is it specifically describing? And why would it be a multi-mapping read when these fields are equal?