tophat --library-type option

Question

[biolab]

Dear all

I am using Tophat2 to map sequencing reads on the genome. However, we only the sequencing data was obtained by Truseq kit. We have no other information about the method of library preparation. So my problem is how to set the --library-type option in tophat mapping? Should I set it to fr-unstranded (unstranded) or fr-firststrand or fr-secondstrand?

Thank you for your suggestions!

Answer 1

[iraun]

You can infer the strandness of your experiment using RSeQC (RNA-seq Quality Control Package) package. Specifically, the script you should use to get information about the strand is the one called: infer_experiment.py. This program is used to "guess" how RNA-seq sequencing were configured, particulary how reads were stranded for strand-specific RNA-seq data, through comparing the "strandness of reads" with the "standness of transcripts".

You can get the package and read all the documentation here

Of course you can extract this information using other methods, this is just a recommendation or an option :-).

Hope it helps.

[michael.ante]

In principle, I'd go for Airan's approach. First you need to align a subset of reads with the fr-unstranded option, check with the RSeQC tools the read orientation and run subsequently Tophat with the computed strandedness.

As far as I know, the library-type option is only adding an XS flag to the reads which is used in Cufflinks afterwards. If you do not want to use Cufflinks, it should be OK just using the fr-unstranded option from the start.

[biolab]

Thank you for your advices，Michael.

[biolab]

Thank you very much, Airan. Your comment is really helpful.