Hello All Please bear with me if this question is silly Can any one tell me the difference between strand specific : yes , no ,reverse in htseq-count I tried with all three as trouble shoot for my practice and i am getting different counts with all. Can any one explain me what does that mean and its significance?
Which setting to use depends on how the sequencing libraries were prepared, though likely one of no or reverse are correct for anything sequenced in the past few years. no is appropriate for datasets that are not strand-specific. reverse is correct of libraries made with a dUTP-based method (this is the most common method). yes is appropriate for older datasets where read #1 indicates the strand of he original fragment sequenced. When in doubt, you can run all 3 on one sample. If the counts from yes and reverse are roughly equal for most genes then the dataset is unstranded (i.e., no is the correct setting). If either yes or reverse produces much higher counts than the other then the appropriate setting is the one giving the higher counts. This will pretty much always be reverse these days. You can also just ask the people who did the library prep. what kit they used or whether it was dUTP-based.
It refers to your RNA-seq protocol. Was that protocol strand specific? I.e., does your read 1 correspond to the sense or antisense of the RNA molecule or random? To determine what it is based on htseq-count results have a look at this post of Devon Ryan A: Does better results in HTSeq-Count mean that the script was run correctly?
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
okay!!!!! thanks Devon Ryan
Hello Devon, During counting the reads I have done one mistake that I counted with unstranded option, however my samples were from reverse stranded. The percentage of mapping was 80% for unstranded and 98% for reverse. But for DEGs I can see there is very less difference. Is it due to the other strand has very less count (5%)?