Hi,
I'm trying to run cuffdiff as part of an RNA-seq project, and it's not working. I've tried leaving it to run overnight 3 times now and every time it starts the job but ends up giving me a tool error for all the files it's trying to create.
I'm using a cuffmerge assembly in my transcript input. Here are other specifications:
- Omit tabular datasets: No
- Generate SQlite: Yes
- Input: BAM files from alignments I did with TopHat (for each, 2 technical replicates were merged with BAM Tools). There are 3 conditions and 6 biological replicates in each.
- Library normalization method: Quartile
- Dispersion estimation method: pooled
- False discovery rate: 0.05
- Min alignment count: 10
- Use multi-read correct: Yes
- Perform bias correction: Yes, "Locally cached", "Mouse (Mus Musculus): mm10"
- Include read group datasets: No
- Include count based output files: No
- Apply length correction: Cufflinks effective length correction
- Set additional parameters for single end reads: No
- Set advanced cuffdiff parameters: No
- Job resource parameters: Use default job resource parameters
Any help would be greatly appreciated.
Mina
Do you have some specifics on the errors you are getting? You can show it using https://gist.github.com/ and pasting the URL here.
Sorry I'm not sure how to use Gist. When I expand the files, there's a blue information notice that says "The job creating this dataset has been resubmitted". Underneath that it says "tool error. An error occurred with this dataset:" but there's no more information given. When I click on the icon for "view or report this error" it says "Remote job server indicated a problem running or monitoring this job" and then it gives me the option to send an error report.
That's really it. That's all the information.
The only thing I can think of, if not something wrong with Cuffdiff itself, it might be an issue with the merged filed I'm using. I've been having problems with Cuffmerge and I discussed this in a different post. There's a problem with assigning a reference genome where the "locally cached" option briefly shows "mm10" and then settles on "no options available". When I created the merged file, there was a brief moment where I though Cuffmerge was working again. It allowed me to choose "locally cached" but there was no drop down menu present to choose a genome. So I just clicked execute and it created a file successfully. But now with the errors I'm getting with Cuffdiff, I'm reconsidering whether the merging actually worked properly. Cuffmerge is still now showing "no options available" under "locally cached"- it's been like this for about a week now, with the exception of the one time I was able to use it without a drop down menu under "locally cached".
So I'm not sure if the problem is necessarily with Cuffdiff or if it's with Cuffmerge, but at least one of those two has not been working for me over the past week.
Hope that gives a bit more context.