Mapping strand-specific data with TopHat

[Comosus]

I have RNA-seq data that I want to map using TopHat. The library prep was done using ScriptSeq v2, and the sequencing using Illumina.

The R1 sequences the sense cDNA from start to finish. The R2 sequences the antisense cDNA from finish to start.

The --library-type option fr-unstranded says that "Reads from the left-most end of the fragment (in transcript coordinates) map to the transcript strand, and the right-most end maps to the opposite strand.", which would seem to be just what I need.

However how does it deal with sense cDNA that has been expressed from the negative strand of the DNA? If an R1 read is just that, then wouldn't the read have to be reversed to map to the genome?

Has anyone dealt with anything similar, and if so, do you know which settings I should use for TopHat?

[Oliviervg]

I have the same problem.
I prepared my library with Scriptseq V2 (directional).

The sequencing was done with Illumina PE 2*76bp.


Can anyone explain more precisely the --library-type option of TopHat ?
I don't understand the difference between "fr-unstranded" and "fr-secondstrand"


What's the best parameters in my case ?

Thank-you.

[Oliviervg]

Any help would be very appreciated !
Thank you !

[Thomas Doktor]

TopHat doesn't have those options, are you sure you're not thinking of Cufflinks?
For the scriptseq v2 kit I would use the fr-secondstrand librarytype since the first read will be the left-most in transcript coordinates.

[Oliviervg]

Thank you for your answer, I m going to use fr-secondstrand.
And, my TopHat 1.4 has those options !

[Thomas Doktor]

You're right, they were just well hidden (and I never had use for anything other than the default).

[harryzs]

Just a reminder, if you did not see this:

http://tophat.cbcb.umd.edu/faq.html