Rnaseq using only coding genes

Question

[maurices]

Hi!!!

If I take from the counts use on DESeq2 only the reads from coding genes and not pseudogenes the results are reliable?

Answer 1

[Devon Ryan]

Sure, most people don't bother analyzing pseudogenes anyway.

[maurices]

Thanks so much!!

If I take all my RNA transcript and annotate using biomart I found many antisense, pseudogenes and coding sequence . If I try to do Unsupervised studies using all rna transcript my cluster are not primary related with the element I try to investigate. Differently If I use only coding genes I found Interesting things!

So there is not a bias on choose only protein coding..

[Devon Ryan]

Maybe there's a bias, maybe not. If these are datasets that weren't created by you then I imagine that library type and/or percentage of poly-A RNA is what's really mucking with what you hope to see (so just looking at protein coding genes helps minimize that).