Question about tophat2 pipeline

Question

[tiago211287]

After using tophat I want to count genes with HTSeq and analyse in R with DEseq. But I have a question. I must input all pair-end fastq files of a experimental group once, or run a tophat for each pair-end fastq file?

Answer 1

[Ashutosh Pandey]

If you have biological replicates, then you should run tophat for each replicate separately. This is because DEseq uses counts from each biological replicate to estimate the dispersion. I don't think HTseq is smart enough to handle library tags in the BAM file and produce counts for each library separately. Hope it makes sense to you.

[tiago211287]

Yes you answered me. Thank you very much for your help.