Extract deletion positions from a SAM file

Question

[egill.richard]

Hello everyone,

I am interested in detecting deletion events in sequencing data, more specifically PacBio data. I searched for a simple solution to detect and extract position of deletion in a SAM file but I could not find anything despite the apparent simplicity of the problem.

In input I have a SAM file that contains in particular : 1) the position of the first base of each read 2) the CIGAR where M stands for a Match, I stands for an insertion and D stands for a deletion

What I want to do is, in each read of my SAM file, getting the start and end position of the deletion.

Input :

pos    CIGAR
1000   200M200D300M

Output :

deletion_start      deletion_end   
1201                1400

I feel that it can be done with a few command lines in python but I am just learning this language. Once I have my table with the deletion I will be independent since I know R much better but for this step I would need your knowledge guys. If there is a tool that does exactly this and that I missed it is even better!

Thank you very much.

Answer 1

[Pierre Lindenbaum]

using bioalcidaejdk. http://lindenb.github.io/jvarkit/BioAlcidaeJdk.html

output is a bed file. (not tested). Basically, the script only loop over the cigar string and print a bed record for each Del

 java -jar dist/bioalcidaejdk.jar -e 'stream().forEach(record->{ if(record.getReadUnmappedFlag()) return; Cigar cigar = record.getCigar(); int ref= record.getStart(); for(final CigarElement ce: cigar) { final CigarOperator op = ce.getOperator(); if(op.equals(CigarOperator.D) || op.equals(CigarOperator.N)) { println(record.getContig()+"\t"+(ref-1)+"\t"+(ref+ce.getLength())+"\t"+record.getReadName()); } if(op.consumesReferenceBases()) ref+= ce.getLength(); } });'  input.bam