How to add read quality scores back into GAM

[BaxW]

GraphAligner is a great tool (thank you). I've noticed that the alignments that GraphAligner outputs never include the base-level quality scores from the aligned reads. I understand that this is expected behavior, and I'm not asking for it to be changed, but I'm wondering if there's a recommended way to pull these scores from the aligned reads and add them into the GAM output by GraphAligner. I want to do this because variant callers I work with require the qual scores, and I imagine others have encountered this issue too.

The obvious answer would be to index the clipped reads to match the gam and then lift over the qual scores, but the --corrected-clipped-out flag outputs reads in fasta, so no qual scores. Any advice or hints would be appreciated.

Thanks!!

[adamnovak]

We're also having this problem, trying to compare GraphAligner to Giraffe as a mapper for calling with DeepVariant (which can't call without quality scores).

GraphAligner also will hard-clip read sequences and just not mention part of the input read in the output GAM sometimes, right? So the algorithm to put the quality scores on at the end would have to deal with that, and work out what offset in the input read the alignment starts/ends at, and take that part of the quality string.