read counts from samtools idxstats

[witty]

Hi,

I want to get the read counts for gene expression analysis. I used "samtools idxstats <bam file>" and got the four columns of "contig ID, read length, #mapped reads and # unmapped reads".

My question is I should use #mapped reads(the third column) or (#mapped reads + # unmapped reads)/2 as the raw reads count?

Thank you in advance!
Victoria

[TiborNagy]

Raw read count = mapped reads + unmapped reads

[witty]

Thank you for the response. Alternatively, this adding value divided by two also makes sense, in this case the read counts are the fragments.
I am not sure which result is from the htseq-count, which is generally used as the input for differential gene expression.

THanks!
Victoria

[dpryan]

htseq-count doesn't count multimappers, so the "total reads" that it would report for something and that reported by samtools idxstats may not be very closely related.