Problems In Transitioning From Tophat To Cuffdiff

Question: Problems In Transitioning From Tophat To Cuffdiff

5.1 years ago by

Elwood Linney • 180 wrote:

After successfully using RNAseq software in Galaxy online for about 10 different datasets to just get gene expression differences between replicates from control versus exposed zebrafish embryos, I am having no luck getting cuffdiff to work with the "moved" Galaxy. I had this problem with histories developed before the move and histories developed after the move. I have had this problem using an order cuffmerge gtf file that worked in the past in Cuffdiff, with a new cuffmerge file developed from cufflinks of the files and by just using a ref file gtf from UCSC. I don't know if this is just some interface problem with a different version of the software that was included with the move, or a reference genome that does not interface with Cuffdiff. It has happened with about 5 different histories. Is anyone else having this problem? And found a solution? Elwood Linney

rna-seq cuffmerge cufflinks • 1.3k views

ADD COMMENT • link •

modified 5.1 years ago by Jennifer Hillman Jackson ♦ 25k • written 5.1 years ago by Elwood Linney • 180

5.1 years ago by

Jeremy Goecks • 2.2k

Jeremy Goecks • 2.2k wrote:

Are you reporting a bug for each failed Cuffdiff run? That's the easiest way for the Galaxy team to help you out. One thing to keep in mind is that, for now, spaces are not allowed in condition names. We'll address this problem soon. Best, J.

ADD COMMENT • link written 5.1 years ago by Jeremy Goecks • 2.2k

5.1 years ago by

Jennifer Hillman Jackson ♦ 25k

United States

Jennifer Hillman Jackson ♦ 25k wrote:

Elwood, Would you please share: 1. an original history that worked (if possible) 2. the workflow you are using and let me know how you generated (did you extract it recently (last week or so)? 3. problem history There are a few known issues in here, some mitigated, one that has to do with extracted workflows from existing older histories that is Cuffdiff specific (actually, a few problems, found another last night). I want to rule those out or at least try to provide feedback. I am at a conference over next several days, so let's use galaxy-bugs as a cc whenever we communicate so nothing is left lingering, plus I know Nate was looking into some of your data. Jeremy may jump also in at any time and help - he is the author of the wrapper. So, in your *reply*, _remove galaxy-user@bx.psu.edu_ and _add galaxy-bugs@bx.psu.edu_ Jen Galaxy team -- Jennifer Hillman-Jackson http://galaxyproject.org

ADD COMMENT • link written 5.1 years ago by Jennifer Hillman Jackson ♦ 25k

Hi Elwood, Jeremy and I took at a look at this. The failures in your history with this message: Error: number of labels must match number of conditions ...are due to a regression in the cuffdiff tool which Jeremy has just fixed. Some characters such as spaces in the condition name were valid prior to the move, but after the move they became invalid. They should now be valid once again. You've got one tophat job running and a cuffdiff job queued, and the rest of your grey datasets are waiting for those two jobs to be finished before they'll be queued (we'll be working on creating a visual distinction for this). I fixed the queued cuffdiff job, which I believe should allow it to run without error (or at least, not the same error). --nate

ADD REPLY • link written 5.1 years ago by Nate Coraor ♦ 3.2k

thank you all, its clearly working now and that is coming just in time, the results of two of these datasets will help us plan our next experiments for capping up a manuscript or two el linney

ADD REPLY • link written 5.1 years ago by Elwood Linney • 180

This is great news, thanks for letting us know. --nate

ADD REPLY • link written 5.1 years ago by Nate Coraor ♦ 3.2k

Please log in to add an answer.

Similar posts • Search »