Problems with Cuffdiff

I’m having some issues with cuffdiff. When I use my reference gtf/gff3 gene notation file, I keep on getting notest and 0 as my output. But, when I look at my cufflinks data output, there is variation seen between my control and experimental samples. When I use my gtf file generated by cufflinks and run cuffdiff, it shows variation between my control and experimental samples. Is my gtf/gff3 file not working or is it a mapping issue? I know that this program has been deprecated due to problems with SQLite, but I can’t use the new differential analysis tools because I don’t have replicates for my data and saving multiple files of the same data doesn’t work. I did bulk RNA-seq which is why I don’t have replicates.

1 Like


The reference annotation GTF dataset probably has a format problem. The most common case is that the the chromosome identifiers are a mismatch between the annotation and the genome mapped against.

Please see this FAQ for how to resolve the problem:

If you are interested in discovery (novel transcripts represented by your read data), run the Cufflinks + reference annotation GTFs through Cuffmerge to generate the GTF for Cuffdiff. Or, you can summarize just known transcripts by using the reference annotation as-is.