"not_aligment" from the output of HTseq-count

[flyingoyster]

What's the meaning of "not_alignment" from the output of HTseq-count? The manual says it is the "reads in the Sam/Bam file without alignment". Does this mean the reads that are not mapped to the reference? or anything else? Thanks!

[dpryan]

Yeah, these are just reads that didn't align. Whether the count for this is 0 or not will depend on the aligner that you're using (some don't put unaligned reads in the output SAM/BAM file, others do).

[flyingoyster]

Assuming that the file I used for input of HTseq-count includes the unaligned reads in the output BAM file, But the number of "no_alignment" reads doesn't match the number of reads not mapped to the reference. Any idea about this? Thanks!

[Simon Anders]

it should match. Maybe it's paired-end? (htseq-count counts read pairs, not reads)

[dpryan]

@flyingoyster: Try using the -o option to see what's going on in your case.

[flyingoyster]

@Simon Anders, Yes, My data is paired-end. If HTseq-count counts read pairs, not reads, that makes sense. Thanks a lot!
@dpryan, I used Galaxy server, so not be able to choose options.

If I want to calculate the number of reads within genes, is it right to substract the number of no_feature reads from the total number of mapped reads? Thanks!