Question: Installing MACS on a local Galaxy
0
gravatar for david.a.morais
3.2 years ago by
Canada
david.a.morais110 wrote:

Hi All,

 

I am trying to install MACS for chip-seq analysis on my local Galaxy.

The devteam  version 1.0.1, revision 6e5fcffb68c9  and the MACS14 owned by 'ryo-tas'  stay in a cloning state and do not install, even after all its dependencies have been successfully installed.

Could someone tell me if this is a known error? Are those repositories just wrappers or they install the MACS python tool itself.

Thank you very much

David

 

toolshed admin macs • 1.4k views
ADD COMMENTlink modified 3.2 years ago • written 3.2 years ago by david.a.morais110
1

Why not using the MACS2 version from the IUC?

ADD REPLYlink written 3.2 years ago by Bjoern Gruening5.1k

Hi Bjoern,
I installed MACS2
After the whole install when I try to test I keep getting this error

Traceback (most recent call last):
  File "/galaxy/tool-dependency/macs2/2.1.0/iuc/package_macs2_2_1_0/4d7ef54c3fa4/bin/macs2", line 559, in <module>
    main()
  File "/galaxy/tool-dependency/macs2/2.1.0/iuc/package_macs2_2_1_0/4d7ef54c3fa4/bin/macs2", line 55, in main
    from MACS2.callpeak import run
ImportError: No module named callpeak

I checked the install dir and the lib/python and all the modules are there.

Any suggestion?
Thanks

ADD REPLYlink modified 3.2 years ago • written 3.2 years ago by david.a.morais110

Hi David, Just jumping in here, Bjoern will know the most about the MAC2 wrapper as he was involved in writing it, but I am wondering if you are using managed dependencies when installing from the Tool Shed into your local. If so, are those dependency package(s) installed correctly when you review them through the Admin tools? Maybe this will help, Jen, Galaxy team

ADD REPLYlink written 3.2 years ago by Jennifer Hillman Jackson25k

Hi,

can you provide us a few more informations? Are other dependencies installable on your system? How do you test these wrappers? You are running it from within Galaxy right?

ADD REPLYlink written 3.2 years ago by Bjoern Gruening5.1k
0
gravatar for Jennifer Hillman Jackson
3.2 years ago by
United States
Jennifer Hillman Jackson25k wrote:

Hello,

It seems like both of these issues are rooted in the same problem - managed dependencies are not installing correctly. 

To answer the original question, the devteam wrapper for v14 should install dependencies including the binary. Make certain that you check this option when installing. If there are issues, the error messages associated with the install should note what is wrong. In particular, is a dependency is missing or is unable to be "checked" for managed install, it means that your galaxy.ini must be modified to specify where managed dependencies should be installed (path to a Galaxy accessible, existing directory).

That said, using the updated tool wrapper is a better choice. So going forward with troubleshooting that is worth it unless you really need to use the older version for some reason (to reproduce a prior analysis, etc).

Thanks, Jen, Galaxy team

ADD COMMENTlink modified 3.2 years ago • written 3.2 years ago by Jennifer Hillman Jackson25k
0
gravatar for david.a.morais
3.2 years ago by
Canada
david.a.morais110 wrote:

Hello,

 

The dependencies were installed without a problem. I manually inspected them.

What seems to be happening is a problem during the export of the PYTHONPATH. The env. variable is properly exported (I echoed $PYTHONPATH from the Galaxy generated .sh file) but the macs2 script does not seems to be able to find the modules even when the PYTHONPATH is pointing to the right location.

To go around this problem I symlinked the '/lib/python/MACS2' dir to the same directory of the macs2 executable '/bin' . Then when the macs2 script does the 'from MACS2.callpeak import run' it finds the module just sitting besides it.

This way the script runs without a problem.

 

Thanks Bjoern and Jenniffer for the help :)

ADD COMMENTlink written 3.2 years ago by david.a.morais110

This is really wired. During job execution what is written to your Galaxy log files. The tool should source the env file before running the script and with this all env variables should be set. Anything special regarding your setup?

ADD REPLYlink written 3.2 years ago by Bjoern Gruening5.1k

HI Bjoern,

I finally found out what was wrong.

There had been a manual installation of MACS2 that messed up the whole $PYTHONPATH.

After I found that out I just needed to clean the old files and do a fresh install of the MACS2 repository.

Thanks a lot for your help, I really appreciated.

 

David

 

ADD REPLYlink written 3.2 years ago by david.a.morais110

Great you figured this out. Keep in mind that we recommend to run Galaxy in a clean virtualenv environment.

ADD REPLYlink written 3.2 years ago by Bjoern Gruening5.1k

Hi Bjoern,

MACS is working great now. However, when the users try to use the option that generates bigwigs (Also generate bigWig file from bedGraph), they get this error

 

/tmp/tmp2ZUQTS/MACS_Pho4_treat_pileup.bdg.clipped is not case-sensitive sorted at line 114573.  Please use "sort -k1,1 -k2,2n" with LC_COLLATE=C,  or bedSort and try again.


I tried to export the env. variable LC_COLLATE=C, but this did not solved the problem.
Do you have any suggestion?

Thanks

ADD REPLYlink written 3.1 years ago by david.a.morais110

Did you export this variable during tool start?

ADD REPLYlink written 3.1 years ago by Bjoern Gruening5.1k

Hi Bjoern,

Yes I added the export directly to the .sh file template DEFAULT_JOB_FILE_TEMPLATE.sh) and it did not help.

 

Cheers

ADD REPLYlink written 3.1 years ago by david.a.morais110

Can you make sure all your input files are properly sorted.

ADD REPLYlink written 3.1 years ago by Bjoern Gruening5.1k

Please also set "LC_ALL=C".

ADD REPLYlink written 3.1 years ago by Bjoern Gruening5.1k

I sorted the bam file with bam sort and also set LC_ALL=C.
Unfortunately no luck.

Cheers

ADD REPLYlink written 3.1 years ago by david.a.morais110
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 177 users visited in the last hour