RNA-seq Infer experiment wrong strandeness information

Question

[abedkurdi10]

Hello everyone,

I have received recently eight RNA-seq samples. We know that the reads are reversely stranded. After aligning those samples to their reference genome, I run infer_experiment on the alignment data just to make sure everything is fine.

Five samples out of eight showed that they are really reversely stranded, however, the remaining three samples showed that the reads are forward stranded. In addition, those three samples show that they contain high percentage of rRNA sequences (using bbduk ~30 percent of the reads).

Removing those reads did not change the result of infer_experiment.

So what could be the reason(s) behind such confusion? Thank you in advance.

[Asaf]

This is weird. Can you extract the reads mapped to forward strand and see what are they? Maybe you have a lot of anti-sense RNA? (long shot, I know). Also, what organism?

[abedkurdi10]

I will have a look on them. Actually, It's mouse data.

Answer 1

[abedkurdi10]

It seems that there is more rRNA sequences. I downloaded rmsk bed file from UCSC and removed all reads covering those Repeating Elements (including and not limited to rRNA, tRNA...), then I run again infer_experiment and everything went well.

[Asaf]

Why did you delete and close the post? It is relevant to the forum, please revert it