Hello,
I am currently analyzing some bulk RNA sequencing experiments for drosophila.
I performed "alignment" and quantification using salmon. The mapping rate looked very good (>80%) and I was happy. I used the -l A flag to automatically infer the library type.
After the alignment, I checked LIBTYPE salmon identified. To my surprise, roughly 50% were classified as ISR and the other half as ISF. This really surprised me, because I know that all samples were prepared using the Illumina TruSeq stranded protocol, so I would expect that all library types are ISR. If I "force" salmon to use -l ISR some samples get a mapping of just 0.5%.
To be sure, I also aligned some samples using STAR followed by RSeQC and they returned the same LIBTYPE as salmon...
Has anyone encountered something similar yet or maybe has an idea why this might have happened?
Any pointers are much appreciated!
Cheers!
