View Single Post
Old 01-30-2014, 02:30 AM   #4
Devon Ryan
Location: Freiburg, Germany

Join Date: Jul 2011
Posts: 3,480

Just browsing through the htseq-count code, it looks like this error will happen if the program things you have paired-end reads and you then hand it a single-end read. You might just feed to unmerged BAM files separately into htseq-count, possibly summing the resulting file in R. Whether you should sum the counts or treat them separately will depend a bit on how they were created. If the paired-end and single-end data were from different library preps of the same sample and you have equivalent data for other samples in your comparison, then I would recommend simply adding "library_type" (i.e., single or paired-end) as a factor in your statistical model (i.e., keep the counts separate). There's an example of that in the DESeq(2) vignette with the pasilla dataset.
dpryan is offline   Reply With Quote