Missing Reads

[christopher-schroeder]

Dear bazam team:

I have noticed that bazam does not output all read pairs. That was quite unexpected for me. Does it have a built-in filter? For example for duplicates? And can you configure it to really output all read pairs?

Best
Christo

[christopher-schroeder]

Oh sorry, never mind. The input bam seems to have missing mates for some reads.
But just to make sure: Bazam outputs all reads and doesn't filter anything (except single reads).

[ssadedin]

Hi, yes you are correct - Bazam will definitely output every read that has a mate and will not filter anything. However currently it does drop single ended reads. Potentially a flag could be added to output them as single ends when streaming into aligners (eg: BWA) that will do "smart" pairing of the reads (as in: if two consecutive reads have the same name they are treated as paired, if they differ they are treated as single ends).