Parsing reads from SAM file by Fwd-Fwd orientation

Question

[ThePresident]

I want to extract mate-pairs that mapped on the same strand, i.e. fwd-fwd and/or rev-rev. As I understand it, this information is encoded on the FLAG field in the SAM file. However, I am not sure which values are used to specifies such reads. One previous post suggested that 113+177 values of the FLAG field specifies reads that both mapped on the reverse strand. However, not sure how fwd-fwd would be encoded?

Also, the number of reads bearing 113 and 177 flag values is not identical in my dataset, but it should be, right?

Otherwise, is there another method of doing this?

Thank you, TP

Answer 1

[Pierre Lindenbaum]

using samjs: https://github.com/lindenb/jvarkit/wiki/SamJS

 java -jar dist/samjs.jar -e '!record.getReadUnmappedFlag() && record.getReadPairedFlag() && !record.getMateUnmappedFlag() && record.getReferenceName().equals(record.getMateReferenceName()) && record.getReadNegativeStrandFlag()==record.getMateNegativeStrandFlag()'  input.bam

[ThePresident]

Thanks for the detailed command line since my java understanding equals 0. I'll give it a shot.