using Cuffdiff without gene and transcript discory

[narges]

Hi all,

I want to use Cuffdiff and I am just interested in differentially expressed genes but not any gene discovery. So I have used the Bam files from tophat output and then using them as the input of cuffdiff. Below is the code I have used but nothing is run and it says there is a command mistake. Is it due to the point that I have not run cufflinks before it can it be due to the point that I have used flag -G before my gtf file just for not interested in discovery of new genes?
Any help would be appreciated.
#!/bin/bash
#$ -cwd
#$ -pe smp 12
module use /app/modulefiles
module add samtools
module add bowtie/2.0.0-beta7
module add tophat/2.0.4
module add cufflinks/2.0.2-beta
cuffdiff -o output -p 12 -G /Genes/genes.gtf -L Male,Female /group1/accepted_hits.bam /group2/accepted_hits.bam

[EGrassi]

Afaik from http://cufflinks.cbcb.umd.edu/manual.html cuffdiff does not have a -G option but just wants the gtf as first argument before the bam files, eg:

cuffdiff -o output -p 12 -L Male,Female /Genes/genes.gtf /group1/accepted_hits.bam /group2/accepted_hits.bam

[xiongdianguang]

I recommended you to read the paper<Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation> .I want it can help you

[xiongdianguang]

by the way. you can try it after you run cufflinks.Because the file produce by cufflinks is necessary when you run cuffdiff.

[narges]

Thank you EGrassi. You are right, now it works

[narges]

by the way. you can try it after you run cufflinks.Because the file produce by cufflinks is necessary when you run cuffdiff.

Thank you so much for the paper. But are you sure that I need the cufflinks results for the cuffdiff? Since in the manual it is written that I can run cuffdiff directly with Bam tophat results.

[EGrassi]

Thank you so much for the paper. But are you sure that I need the cufflinks results for the cuffdiff? Since in the manual it is written that I can run cuffdiff directly with Bam tophat results.

(For nothing!)
If you do not want to work with new possible isoforms suggested by you RNAseq data (produced by cufflinks) you can just use a known GTF as the annotation used to align your reads with tophat and then for cuffdiff directly.

[glados]

You do not need to run cufflinks before cuffdiff! Especially not if you're not interested in new genes/isoforms. Then just use your annotation gtf and the bam files. They even mention it in the manual.