Call TopHat on Bowtie2 output?

[tokbio]

Does TopHat have a switch to work on output from Bowtie2? This would be useful for flexibility in modular pipeline/batching, and also in view of apparent issues in dealing with multiple hits (not sure of the current status on that). On a related note, it appears TopHat calls bowtie once, but are there instances where it might call it multiple times?

[tokbio]

Hmm, the junction finding step probably prevents any simple bowtie > tophat modularity here.

[Dario1984]

You want to run a mapping program on the output of another mapping program ? It makes no sense.

[bruce01]

Run tophat2 using the -b flag to specify how bowtie2 runs under it. Read the *tophat2* manual.

[tokbio]

Thanks for the replies. It seems there are two places one can pre-compute:
1) bowtie2 to index the reference genome (of course)
2) tophat -G to index the transcriptome
I.e. indices can be pre-computed, which minimizes calls to bowtie2, overhead if the tophat algorithm changes and legacy experiments are re-run, etc., as long as one is happy with the reference genome and transcriptome, of course.