How to join the different Gene ID from edgeR/DESeq and Cuffdiff?

[super0925]

Hi All I am a rookie in RNA-seq.
I found a problem that gene ID output from Cuffdiff and egdeR/DESeq are different so that I cannot find the common DE genes.
My two pipelines:
Tophat-Cufflinks-Cuffmerge-Cuffdiff
Tophat-edgeR/DESeq
I use the data set with the Drosophila melanogaster genome.
The edgeR output gene ID (e.g. FBgn0000370,FBgn0000500,…)
But the Cuffdiff output gene ID is something like (XLOC_000028,XLOC_000038,…) and gene symbol (e.g.,KH1,RpLP1,…).
I found that the Cuffdiff 's gene ID is corresponding to merged.gtf (by Cuffmerge) but egdeR/DESeq's gene ID is corresponding to genes.gtf (by reference genome)
Could you please recommend any software or command to translate them automatically? Could you have command line for me to use to uniform or join them?
Thank you!

[dpryan]

Just perform the counting on the GTF file generated by the cufflinks pipeline. Then the IDs can be the same.

[super0925]

Just perform the counting on the GTF file generated by the cufflinks pipeline. Then the IDs can be the same.

What .gtf? Do you mean transcripts.gtf generated by Cufflinks or merged.gtf generated by Cuffmerge?
I found that the Cuffdiff 's gene ID is corresponding to merged.gtf but egdeR/DESeq's gene ID is corresponding to genes.gtf
However, from the Manual I saw that the edgeR/DESeq should use genes.gtf(from reference genome) to generate the counts table.
Could I use merged.gtf as well？

Thank you very much!

[dpryan]

merged.gtf

If you're going to make a comparison, that'd be the most apples-to-apples. Neither the DESeq(2) nor edgeR manuals will have anything to say on this matter, since it's out of their purview.

[super0925]

Thank　you!
I will try to generate the counts table based on the HTSeq-count and merged.gtf

[super0925]

merged.gtf

If you're going to make a comparison, that'd be the most apples-to-apples. Neither the DESeq(2) nor edgeR manuals will have anything to say on this matter, since it's out of their purview.

Hi Dpryan I also find a post discussing about if we could use merged.gtf as HTseq-count to generate the counts table

http://www.researchgate.net/post/What_is_your_favorite_DEG_test_for_RNA-seq_data

Please see first two posts (i.e. "Popular Answers")
It seems that your opinion is in agreement with Juan Steibel (the first one) , but my previous worry is that Fabrice Chatonnet (the second one) said.

[dpryan]

Fabrice's reply didn't raise any concerns regarding this. Fabrice was referring to people sometimes trying to put RPKM/FPKM values into edgeR/DESeq, which is going to lead to bad results. That wouldn't be done here, since you're just using the same annotation file (alternatively, you could also just have cufflinks not find novel isoforms or splice junctions and only follow the annotation file you give it). You would just use that with htseq-count.

[super0925]

Fabrice's reply didn't raise any concerns regarding this. Fabrice was referring to people sometimes trying to put RPKM/FPKM values into edgeR/DESeq, which is going to lead to bad results. That wouldn't be done here, since you're just using the same annotation file (alternatively, you could also just have cufflinks not find novel isoforms or splice junctions and only follow the annotation file you give it). You would just use that with htseq-count.

Got it, Cheers