Count difference by htseq-cout and samtools

[DZhang]

Hi,

I used bowtie to align a set of pair-end reads to the reference sequence and I used two programs to count the reads:

1) samtools - "samtools view -c -q1 my.sorted.bam chr_ID"
2) htseq-count - "htseq-count -s no -t CDS -a 1 my.sorted.sam my.gff"

One interesting thing about these counts is that the counts by bwa is exactly twice of those by htseq. Can anybody help explain?

Thank you,
Douglas

[Simon Anders]

"by bwa"? You mean "by samtools", I suppose.

htseq-counts counts the two reads from the two ends of a paired-end fragment as one count. Samtools might count them separately, giving two counts.

[DZhang]

Yes, it should be samtools.