Cufflinks and CuffDiff

[snape_ar]

I have 2 RNA-Seq biological replicates and I ran Tophat and then Cufflinks on it separately for both samples. The outputs of Cufflinks for each replicates are:

Replicate 1:
1. transcript.gtf file
2. genes.fpkm_tracking
3. isoform.fpkm_tracking

Replicate 2:
1. transcript.gtf file
2. genes.fpkm_tracking
3. isoform.fpkm_tracking

I would like to compare the differential expression between the two biological replicates by running CuffDiff. So, my question is when I am running the following command:

cuffdiff [options]* <transcripts.gtf> <sample1_replicate1.sam[,...,sample1_replicateM]> <sample2_replicate1.sam[,...,sample2_replicateM.sam]>... [sampleN.sam_replicate1.sam[,...,sample2_replicateM.sam]]

What should be my transcripts.gtf file ? Is it the combination of the two transcripts.gtf files or is it of replicate 1/2 ?

Thanks.

[Nicolas]

I used the annotation files used in Tophat/Cufflinks but your question makes me realize that this is confusing. The manual page says:
"<transcripts.(gtf/gff)> A transcript annotation file produced by cufflinks, cuffcompare, or other source."

but the How to start page (http://cufflinks.cbcb.umd.edu/tutori...l#differential) says:
cuffdiff annotation.gtf mock_rep1.bam,mock_rep2.bam \
knockdown_rep1.bam,knockdown_rep2.bam

I had good results with the annotation file, but I probably missed the novel transcripts. If that's what you're after, you should merge the 2 transcript.gtf and run cuffdiff with it.

[jbrwn]

what nicolas said. use cuffmerge to merge the two *.gtf files from cufflinks output.