Filter our reads in sam file less than specific length

Question

[daewowo]

I have a sam file with reads that are soft and hard clipped. I want to filter out reads less than 32nt including soft and hard clipping, ie the total read length should be less than 32nt irrespective of whether any part of the read aligns to a reference.

Any ideas on how to do this?

Answer 1

[Pierre Lindenbaum]

using http://lindenb.github.io/jvarkit/SamJdk.html

java -jar dist/samjdk.jar -e 'final int len=32; if(record.getReadLength()>=len) return true; if(record.getReadUnmappedFlag()) return false; Cigar c = record.getCigar(); if(c==null || c.isEmpty()) return false; int n=0; for(CigarElement ce: c) { CigarOperator op = ce.getOperator(); if(op.isClipping() || op.consumesReadBases()) n+=ce.getLength(); } return n>=len;' in.bam