Filtering RNA-seq data

[swarbre]

Hi,

I have something of a newbie question regarding filtering RNA-seq data. I've downloaded some RNA-seq data from SRA in fastq format and plan to align this to the genome using tophat.

Is it necessary/standard to filter this prior to generating the mappings, I guess the mapping itself is acting as a kind of filter as poor quality data presumably wont be aligned is this sufficient. It would be easy to filter out sequences with stretches of Ns is that typically done. What about using the quality info in the fastq file should I be using this for some filtering prior to mapping?

I assume the data deposited in SRA generally represent some form of quality filtered set and not totally raw data is that correct?

Any pointers or links to info greatly appreciated.

dav

[hypatia]

Hi, I have the same question as you. Did you figure it out?

[jdanderson]

Hello Swarbre and Hypatia,

I think Bowtie/Tophat may do some quality trimming in its processing, but if you want to have some more control of the quality of reads that pass you can use FASTX Toolkit:

http://hannonlab.cshl.edu/fastx_toolkit/index.html

I just finished using its Barcode Splitter module with good results, and its fairly user friendly. ALso there are some useful tools on Galaxy, by Penn State:

http://main.g2.bx.psu.edu/

Hope this helps.

Regards,
Johnathon

[hypatia]

Thanks jdanderson,

I am using Galaxy and I want to follow the quality steps suggested in Wilhelm Nature Protocol papaer for Illumina, but I can not figure it out in Galaxy. (I am new to this world, statistician a lot of other omics data of experience)

[jdanderson]

Hello Hypatia,

So I just gave the Wilhelm paper (Defining transcribed regions using RNA-seq) a quick look over and it appears that they are talking about the quality filtration that occurs in Gerald, which is part of Illumina's CASAVA program package. If you use Gerald you should get four different output files: export.txt, extended.txt, sequence.txt, sorted.txt.

The sequence.txt file is the one that has the quality filtering already done on it by Gerald (and most people seem to use this file). Also, I believe there is a video tutorial on trimming quality reads in the Galaxy site (although i can't remember how detailed it was).

As for specific SRA files, you may want to look over the protocol for the reformatting procedure as this may provide some insight as to which file type is suggested people start with. I have not looked over their protocol yet, but it should be on their site.

Hope this helps.

Regards,
Johnathon