I ran bowtie2, using bulk paired RNA-seq data, with the forward and reverse strand flipped. In other words, the reverse-strand fastaq file was using the -1
argument and the forward-strand fastaq file was using the -2
arguement. Overall, there was little difference between the alignment rate (e.g. 85.38% versus 85.17%).
This struct me as odd. Can anyone explain this behavior?
Thank you!
Did you provide any other parameter related to read orientation ?
If you know you flipped the data then re-do the analysis. It will cause issues with downstream analysis.
Aligners will try and reverse complement the reads to see which strand they map to. Did you check on the insert size stats? In your alignment files, reads would be marked as in wrong orientation (i.e. not properly paired). TLEN may also be marked as abnormally large or negative.