Star alignments and samtools flagstat

[David Lyon]

I ran some paired-end RNAseq alignments using STAR (default options) and then I ran samtools flagstat on the sorted bam files I always get
'0 + 0 duplicates'

All the other stats from flagstat appears consistent but I cant imagine with several million PE reads I get no duplicates.

Any Ideas

[atcghelix]

Hi David,

Flagstat just counts the SAM flags, so if you haven't marked reads as duplicates (generally through a tool such as picard MarkDuplicates), then you won't see any.

Whether or not you want to mark/remove duplicates in RNA-seq is another matter. I think for many RNA-seq applications, such as differential expression with high depths, it is not desirable to do so.

[David Lyon]

Thanks a lot for your help.

Cheers

Dave