I aligned strand-specific RNA-seq samples using tophat2 with default settings and then viewed the resulting "accepted_hits.bam" file in IGV. All the reads aligned antisense to genes. For example, at GAPDH (+ strand gene) all the reads align on the - strand (arrow on left of the read). For ACTB (- strand gene) all the reads align to the + strand (arrow on right of the read).
I figured this is because I used a dUTP first strand library prep protocol and that I would need to specify the prep protocol when aligning. So I ran tophat2 again, this time including the "--library-type fr-firststrand" parameter but the result comes out the same. When viewing the bam files in IGV, the reads are still antisense to genes.
Does anybody have advice on how to fix this?
This also doesn't seem to be an issue with the sequence somehow getting flipped mistakenly during a processing step because if I search for several stretches of GAPDH sequence in the raw fastq file, I won't find any reads having that sequence, but I will find many reads containing the reverse complement of that sequence.
So basically, the sequenced read represents the reverse complement of the actual RNA molecule. Is there a way to specify this during alignment so that the resulting bam shows the correct orientation (ie. reverse complement of the sequenced read)?