Htseq Count Sam

Question

[Varun Gupta]

Hi everyone

I want to use htseq-count to count the number of reads for the features i have but htseq-count says:

If you have paired-end data, you have to sort the SAM file by read name first.

How can i do that if i don't want to use msort.

Regards V

Answer 1

[Linxe]

Use samtools for sorting: samtools sort -n file.bam filesortedbyreadname

[Malachi Griffith]

Does 'samtools sort -n' properly handle paired end data with multiple alignments per read (e.g. rna-seq reads aligned by TopHat)?

[roll]

Hi Malachi, i am having exactly the same problem. Did you sort out how to deal with this?

[Devon Ryan]

For htseq-count it won't actually matter as multimapping reads will be ignored in any case (htseq-count looks at the NH:i: auxiliary tag).

[venks]

You have mentioned that SAM file has to be sorted but in the answer you have mentioned a BAM file. ??? Is that a TYPO ???

Answer 2

[Irsan]

You can also use SortSam.jar from picardtools

java -jar /path_to_folder_picardtools/SortSam.jar INPUT=yourfile.sam OUTPUT=readSorted.sam SORT_ORDER=queryname VALIDATION_STRINGENCY=lenient

Answer 3

[sethugunja]

I have the same question, I have sorted the .bam file and convert it to .sam file : didnt work for ht seq, then I have sorted the .sam file and given for ht seq, didnt work..!! Do you have any other suggestions for sorting the sam file (paired end data) by read name?

[Devon Ryan]

Firstly, try starting a new thread rather than just answering an almost year-old question with another question. Secondly, we can't help you if you just says "does not work" without providing any details. We're good, but we don't read minds.