Bowtie/Tophat SAM format detail question

[traeki]

Hi, there,

I'm trying to get a clear picture of the semantics of the cigar field for bowtie SAM output.

My current impression is that the cigar field is listed relative to the plus-strand genome encoding, so that if I have

genome:
***   **       <- read alignment
ATGCCCGTAA

read: 
AC___CAT

I'd see the cigar string in the resulting alignment aligning at position 0 with cigar 3M3N2M, even though the alignment actually runs from right to left (i.e. the 16 flag is set). Is this correct? Is this documented anywhere? Is this the intended behavior? I think that's consistent with the existing behavior for the MD:Z: string, I just want to find out if it's something that's going to change under me in future versions, and make sure I haven't misunderstood.

Incidentally, the bowtie-format alignments' mismatch fields are NOT consistent with this, or even with themselves (the locations are strand-relative, but the reported base transitions are not!!), so I'm pretty sure that's a bug.

But I'm dealing with SAM output, so I don't really care about the bowtie format. I just want to know the intended semantics for the cigar field in greater detail.

Thanks,

-John