STAR mapping statistics don't add up with samtools count

Question

[Zee_S]

Hello Biostars community,

I have PE samples mapped with STAR and i want to count the number of uniquely mapped reads using samtools. i did the following (using the Alex Dobin's advice in this link: https://groups.google.com/forum/#!topic/rna-star/pQbqeQd0lNY) :

SE: $ samtools view -c -q 255 -F 0x2 Aligned.out.bam
PE: $ samtools view -c -q 255 -f 0x2 Aligned.out.bam

however the SE +PE/2 read count is more the number of uniquely mapped reads in my STAR log.final.out file.

To be more specific, my calculations with samtools add up to :

34 (SE) + 70579527 /2 (PE) = 35289797.5 reads.

The star total unique mappers is 35289751 reads (ie. 46.5 reads less than the samtools count) .

I have tried with a different sample, and still get 46.5 reads more with samtools.

Does anyone know why I'm getting this discrepancy?

Thank you for your help.

[Mark]

Can you provide the STAR statistics log file?

I believe STAR excludes SE reads. The remaining 10 or reads are probably not included because they're either multimapped or something else.

I wouldn't worry about it, ~46 reads is very little to worry about!