Mapping Ribo-Seq and mRNA-Seq from the same project

[Assa Yeroslaviz]

Hi,

I have a project with both RNA-Seq and Ribo-Seq. I would like to know if it makes more sense to map them with similar parameters or should I use short-reads specific parameters for the Ribo-Seq data?

After trimming and preprocessing the Ribo-Seq data I have reads in varying lengths between 12-63, while the RNA-Seq is all 75nt long. 

In the genome indexing step:

Do I need to create two different indexed genomes for the two data sets?

What would you use as a value for sjdbOverhang = ?

In the mapping step:

Does it make sense to include in the Ribo-Seq mapping also such parameters as:

--winAnchorMultimapNmax 200

--seedSearchStartLmax  15

--outFilterMultimapNmax  8

outFilterMismatchNmax 4 

thanks in advance

Assa

[Alexander Dobin]

Hi Assa,

sorry for the delayed reply.

with a highly variable read length for ribo-seq, it's hard to select an ideal value for sjdbOverhang.

I think simply using 74 for both RNA-seq and Ribo-seq will work fine.

Same is true of other parameters. If you are serious about optimizing the parameters, I would recommend selecting a metric of merit (e.g. % of mapped reads) and varying the parameters to optimize it.