Filter out rRNA in variousspecies

[TofuKaj]

Dear great helpers,

I'm new in RNA-seq analysis and I am confused with the idea about filtering the rRNA contaminated sequences. Would you please kindly clarify me about these isssues?

1. I would like to known if it is routine work to filter out rRNA contamination in coding RNA-seq data and when should we do it.

2. Should we filter rRNA before or after alignment (some say we should perform filtering with bowtie before actual alignment with the prefered programs)

3. I'm working with various species, and I'm totally lost here what is the most common ways to perform such filtering. Some sample show me how to perform rRNA filtering in human or mouse but rarely mention about other species (I generally use Ensembl genome in work).

Thank you very much in advances for your great help,
Kaj

[diego diaz]

Hi,

There are two ways to purify your RNA-seq data, Ribo-Zero and Poly(A). The first one only filters out the ribosomal RNA, keeping coding and small non coding RNA. Focused on valuable RNA species leading to discovery of relevant genes, splice variants and isoforms. The second one is a different approximation, you deplete coding transcript by Poly(A), keeping the depleted RNA.

The way to check RNA contamination is using a Bioanalyzer chip, in the wet lab. But always is good to be sure that your data isn't contaminated. I think the best way to check that is aligning your reads to SILVA, a RNA database, and discard those reads that have a significant alignment.

http://www.arb-silva.de/

[GenoMax]

@Kaj - Since 80+% of RNA in a mammalian cell is rRNA (probably similar numbers for others) it is best to remove it before library prep (unless you are working with rRNA). You could do it informatically (if that was the question you were asking) but then you are going to have only a small fraction of your reads (5% or less) representing mRNA.

[TofuKaj]

Thank you very much