I am using hisat2 (v2.4.0) for aligning my stranded RNAseq data and just stumbled upon a downstream analysis issue of which reads come from forward and which reads come from reverse strand. This is because I got an variation in a position present in overlapping exons. One exon is present in the negative strand and another is present in the positive strand (both have different genes). Now, I can't technically tell which where the variation is from as it can come from any of the two transcripts. Correct me if I am wrong.
I also know that variants are reported only from the forward strand. So, does that mean that variation present in genes present in the reverse strand of the genome is not reported?
My knowledge tells me that in the stranded truseq protocol, the reverse strand is sequenced. Unfortunately, while running hisat2, I did not put --rna-strandness as one of the options and only used --dta. Hence, not all reads have the XS tag. XS attribute tag: '+' means a read belongs to a transcript on '+' strand of genome. '-' means a read belongs to a transcript on '-' strand of genome. But some do have. I am not aware how hisat2 did that. Please let know if it is a bug. By default, hisat2 alignment is unstranded.
Should I have cared to use --rna-strandness option so that the current overlapping exons issue could have been avoided? Can somebody enlighten me how to tackle this issue. Please let know if I have not explained my issue articulately enough. I would be grateful if someone can clear these basic doubts (may not be obvious for many people)