Increase FragPipe’s allocated memory

sxhao · July 16, 2026, 4:32pm

Hello, I am writing to ask whether it is possible to increase the memory allocation for FragPipe. I have around 3 million peptides. Even after reducing the number of samples and adjusting other parameters, I am still unable to get output results from FragPipe. I would appreciate it greatly if you could offer some suggestions

jennaj · July 16, 2026, 5:08pm

Welcome @sxhao

Thank for your query!

I double checked and FragPipe is already allocated the largest memory allocation at the public UseGalaxy* servers.

github.com/galaxyproject/tpv-shared-database

tools.yml

c051fd8e3


      
              MONO_GC_PARAMS: max-heap-size={int(mem)}g
          toolshed.g2.bx.psu.edu/repos/galaxyp/fragpipe/fragpipe/.*:
            cores: 12
            mem: 256

You could double check that UseGalaxy.org.au is not reducing this by testing at UseGalaxy.eu and UseGalaxy.org with your specific data, but from what I know, all are using the same allocations. The AU team will be here later today and can comment more! You could also write in directly to them (find the email under the ? icon in the server’s masthead or on the homepage).

Then, if you are certain the error is due to the size or complexity of the data (as opposed to a file format/content issue), you’ll likely need to simply further to get this work to process.

We have an example of the manipulations in this tutorial.

Hands-on: Neoantigen 2: Database merge and FragPipe discovery / Neoantigen 2: Database merge and FragPipe discovery / Proteomics

Does this help to frame the situation? Would you like some help confirming the inputs? Sharing your history is usually how to get more feedback about this.

Let us know how this turns out!

sxhao · July 17, 2026, 2:16am

When executing FragPipe, the generated PSM, ion and peptide output files are empty. After reviewing the error logs, I observe the error shown below:

jennaj · July 20, 2026, 5:24pm

HI @sxhao

Yes, that message is reporting the job is too large for the public clusters.

You could write in directly to the AU administrators to see what they think? If you are an AU national researcher maybe more can be done. As far as I know, the current allocation is maximum available for the public access I’ve seen (across tools!) so there isn’t a request I can help you to make.

The final option (apart from subsetting the data) is to move to a private cluster that can handle this work. The AnVIL version of Galaxy is one choice.

Wish I had a better answer!