using tophat to find differential splices of ONE gene

[Fedster]

I have the CEU RNA seq fastq data from here

http://jungle.unige.ch/rnaseq_CEU60/

I would like to use tophat to see if I can find more than one splice variants for one gene (actually two, the two genes are quite similar, but on different chromosomes).

I have a nice small fasta file with a number of possible predicted splice variants for the gene(s); I also have the hg18 index sequence for bowtie.

It is not quite clear to me what's the correct step now:

tophat /path/to/h_sapiens reads1.fq,reads2.fq,reads3.fq

does suspiciously look like it will happily ignore the fact I only care for one gene, rather than the whole hg18 genome.

I am on a cluster, so I cannot be interactive and I cannot test stuff easily just by trying (and I'm running out of disk space!). How do I tell tophat to just align stuff to my gene in the small fasta file? do I have to run bowtie on the fasta file first somehow?

[sjm]

I think as a first step, you will indeed need to have bowtie build a database from your fasta file.

bowtie-build /path/to/fasta name_of_database

Then have tophat align your reads to that database:

tophat name_of_database reads1.fq,reads2.fq (etc.)