This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Soft-clipping read ends based on read group

Is it possible to clip (soft-clip preferably) n (for example, 3) nucleotides from both ends of reads in a bam file, but only for the reads with a certain defined read group?

I have merged bams for ancient DNA samples and the bams used for merging come from both UDG-treated and non-UDG-treated libraries. I would like to clip 3 nts from the ends of only the reads coming from non-UDG-treated libraries. These are marked by a label in Read Group: @RG LB:**dmg

So as a result I would like to have the same bam output but with the "dmg" labelled reads clipped

I hoped for gatk ClipReads or perhaps trimBam function of bamUtil to allow this but I haven't really found any clue as to whether and how that could be done.

Any tips appreciated.

Cheers, Martyna

ancient molecule at bam dna rg ends clipping damage
• 2,216 views • link
updated by chenl • chenl • written by Martyna • Martyna

Hi,

Can you please share how did you clip both ends of the reads with ClipReads? Clipping form the start with -CT "1-3" for instance works just fine, but I can't figure out how to clip the opposite end as well (with reads having different lengths).

Thanks!

• 0 views • link
written by chenl • chenl

2 answers

split the bam per read group (samtools view --read-group xxx) , clip the read with your solution above, merge the groups (samtools merge).

• 0 views • link
written by Pierre Lindenbaum Pierre Lindenbaum

I think you could use a multistep approach where you split your SAM file by readgroups with

samtools split

trim the relevant sample with either of the tools ClipRead or bamUtil trim (softclip is supported):

then merge the BAM files back together with

samtools merge
• 0 views • link
written by Istvan Albert Istvan Albert

@Pierre Lindenbaum, @Istvan Albert, thanks a lot! My brain has somehow fixed on trying to do this in a single step.

• 0 views • link
written by Martyna • Martyna

Log in to answer this question.

More posts like this