Why do ribo-seq sequencing data have only a small proportion of uniquely mapped reads?

Question

[Cacau]

I downloaded several published ribo-seq data and mapped them with STAR. However, only 15% ~ 50% of them are uniquely mapped reads. The reads are short, the length being about 35bp long after trimming. Is this the reason? Or might be because many rRNA reads are included?

Answer 1

[mark.ziemann]

You're correct in that it could just be short length or rRNA. It is relatively easy to check the following:

1. Screen rRNA content http://genomespot.blogspot.com.au/2015/08/screen-for-rrna-contamination-in-rna.html
2. Simulate cDNA reads of the different lengths and determine the proportion that map uniquely http://genomespot.blogspot.com.au/2014/10/rna-seq-aligner-accuracy-tested-with.html

[Cacau]

That's very helpful! Thank you.

Answer 2

[mark.ziemann]

You're correct in that it could just be short length or rRNA. It is relatively easy to check the following:

1. Screen rRNA content http://genomespot.blogspot.com.au/2015/08/screen-for-rrna-contamination-in-rna.html
2. Simulate cDNA reads of the different lengths and determine the proportion that map uniquely http://genomespot.blogspot.com.au/2014/10/rna-seq-aligner-accuracy-tested-with.html