How to analyze RNA-Seq with 3' bias

[aquleaf]

Hi All.

I'm trying to identify differentially expressed genes between 2 groups. However, here is the problem. My RNA-Seqs are lack of 5' alignments. I manually checked several genes which were known to be differentially expressed by viewing the alignment files. Their expression changes in RNA-Seq were consistent with the knowledge. So I think I still could use this dataset.

My question is
is there any software of differential expression analysis suitable for such dataset lack of 5' alignments?

Wish your help! Thanks a bunch!

Best

[dariober]

Hi All.

My question is
is there any software of differential expression analysis suitable for such dataset lack of 5' alignments?

Hi,

The short answer is yes. Methods like those implemented in edgeR or DESeq (Bioconductor) compare the number of reads overallapping genes in two or more conditions to assess differential expression and they don't care *where* in the gene the reads are. So technically you are fine.

A drop of coverage towards the transcript ends (more towards the 5' ends) is normal. However, if your underrepresentation of 5'ends is very marked without a good reason I would be more worried about the reliability of your libraries before going to the differential expression analysis.

All the best
Dario