how to analyze multiple samples using tophat

[nunu_ping]

I have RNA-seq data from different experiments. How can I analyze multiple samples together using tophat? I tried command as below

tophat -o output_dir --solexa1.3-quals --num-threads 12 hg19 rna_seq/raw_data/cancer_sequence.txt,rna_seq/raw_data/control_sequence.txt"

It seems the result only has one .accepted_hits.bam file and one junctions.bed file. I thought there should have different outputs for each sample. Should I run tophat for each sample separately??

Thanks!

[nunu_ping]

One more question here. I plan to use Cufflinks that uses tophat outputs. It seems Cufflinks only accept one input SAM file. Should I first parse the accepted_hits.bam to get SAM files for each sample and then input them separately to Cufflinks for expression quantification?

What is the good way to use the tophat-cufflinks pipeline?

Much appreciated!!!

[Aurelien Mazurie]

I am learning how to use TopHat, and I do have exactly the same question. Do you happen to know the answer?

Aurelien

[zack80.liu]

Me too. I ran into the same problem... Any idea?

[frozenlyse]

It's not really a straightforward problem, especially since the documentation for tophat/cufflinks is quite sparse - but what I do at the moment is

• tophat each sample separately (pooling technical replicate lanes at this step)
• cufflinks each sample separately
• cuffcompare all the samples, using the -G option providing an ensembl gtf of gene annotations to create a new gtf containing ensembl genes and any novel transcripts discovered by tophat/cufflinks
• cuffdiff samples against each other using the new gtf created in the previous step (pooling biological replicates at this step)

Any suggestions?

[Bioboatr]

Hi all. definitely search around on this site. there are a lot useful answers if you look hard enough, but it does take some patience. I can suggest at least a few threads (below). Good luck!

Jamie

http://seqanswers.com/forums/showthread.php?t=3967

http://seqanswers.com/forums/showthread.php?t=5180

http://seqanswers.com/forums/showthread.php?t=9677