clipOverlap from BamUtil modifies CIGAR string => error in GATK

Question

[Juke34]

Hi,

I have a problem with gatk remove_duplicates that throws the error No real operator (M|I|D|N) in CIGAR after using clipOverlap from BamUtil. Without the clipOverlap step I do not have the problem within GATK. So the problem is obvisouly introduced by clipOverlap. Does Someone know how to fix that?

My command is as follow: bam clipOverlap --storeOrig CG --poolSize 50000000 --in $bam --out ${bam}_ovClipped.bam

Answer 1

[lechu]

It's only a guess, but maybe deduplicating before clipping would solve the problem...?

[Juke34]

I was removing the duplicates before this step if it is what you mean, with MarkDuplicates

[lechu]

Then you should rephrase your question, because not matter how many times I read it, it seems to imply the opposite. Maybe you are using wrong SAM version (1.3 vs 1.4)? Since the error message suggest sth is wrong with you CIGARs, maybe you can update you question with a short excerpt from your BAM files before and after being processed with BamUtils.