Cloudman - Galaxy - TopHat2 error

Question: Cloudman - Galaxy - TopHat2 error

4.6 years ago by

mg1655 • 20

United States

mg1655 • 20 wrote:

Hi all,

I'm going through the RNAseq tutorial posted here:

https://usegalaxy.org/u/jeremy/p/galaxy-rna-seq-analysis-exercise

Data upload and Fastqc goes well (nice job!); however, I run into a problem with the tophat step.

Tophat errors out with the paired end setting.

Using the reads in a single end mode works (I guess because the closure-search is irrelevant for SE reads)

An error occurred with this dataset:TopHat v2.0.9 tophat -p 4 -r 110 -a 8 -m 0 -i 70 -I 500000 -g 20 --library-type fr-unstranded --max-insertion-length 3 --max-deletion-length 3 --coverage-search --min-coverage-intron 50 --max-coverage-intron 20000 --closure-search --min-closure-exon 50

Dataset 27: Tophat for Illumina on data 14 and data 19: accepted_hits

Tool execution generated the following error message:

Error in tophat:
tophat: option --closure-search not recognized
    for detailed help see http://tophat.cbcb.umd.edu/manual.html

Thanks!

tophat cloudman • 1.2k views

ADD COMMENT • link •

modified 4.6 years ago by Dannon Baker ♦ 3.7k • written 4.6 years ago by mg1655 • 20

4.6 years ago by

Jennifer Hillman Jackson ♦ 25k

United States

Jennifer Hillman Jackson ♦ 25k wrote:

Hello,

There is most likely a dependency mismatch between the installed repository and the underlying binary.

Specifically, Tophat (v1) includes this option, but Tophat2 does not. The Galaxy wrapper and tool form that are associated with Tophat (v1) also include this option, from this repository: http://toolshed.g2.bx.psu.edu/view/devteam/tophat

When you are executing the job, the binary for Tophat2 called by the wrapper for Tophat (v1) is causing the conflict. When installed manually, Tophat2 helpfully includes a symbolic link "tophat -> tophat2".

All of this is a best guess at what is going on, but it has come up before. You either have just these mismatched pieces installed - or both versions of the binaries and wrappers installed. Your genome indexes are most likely a match for the tool - each require a different format - get those from our rysnc server or generate your own:

https://wiki.galaxyproject.org/Admin/UseGalaxyRsync (download)
https://wiki.galaxyproject.org/Admin/DataPreparation (create)
https://wiki.galaxyproject.org/Admin/Tools/DataManagers (new create option)

Going forward, just make sure the tool paths are correct, remove any symbolic links that interfere, use the correct indexes for the tool version, and see if the problem resolves. Next time, if you install directly from the Tool Shed with dependencies, the managed dependencies install process will take care of the binaries for you (or should!). Data you will still need to manage separately.

Best, Jen, Galaxy team

ADD COMMENT • link written 4.6 years ago by Jennifer Hillman Jackson ♦ 25k

Thanks for the suggestions.

I launched my instance using usegalaxy, so there is zero configuration done on my end.

I will try the genome index angle.

ADD REPLY • link written 4.6 years ago by mg1655 • 20

You used CloudLaunch? There is still likely a wrapper-binary mismatch (from the information given). We are looking to the ami config and will get back to you shortly.

ADD REPLY • link written 4.6 years ago by Jennifer Hillman Jackson ♦ 25k

4.6 years ago by

Dannon Baker ♦ 3.7k

United States

Dannon Baker ♦ 3.7k wrote:

The issue you're finding with the Cloudman instance is indeed that while some options are compatible between versions (I was able to successfully map and continue with the rna-seq exercise with the default parameters) -- this option is not compatible. We're actually updating the cloud deployment over the next week or so to contain both tophat1 and tophat2 options and the relevant bits required to run both individually (among other updates).

One solution for you right now, though would be to install (on your cloud instance) the most recent tophat2 wrapper from the toolshed and use that -- this is just a few clicks in the admin interface.

ADD COMMENT • link written 4.6 years ago by Dannon Baker ♦ 3.7k

Please log in to add an answer.

Similar posts • Search »