Question: MACS2 bdgdiff and Diffbind problem - job was terminated because it used more memory than it was allocated
gravatar for johanvn
10 weeks ago by
johanvn10 wrote:

Hi all. I'm having som trouble running both MACS2 bgddiff and Diffbind. It keeps returning the error: "This job was terminated because it used more memory than it was allocated."

I have 2 ChIP samples with 2 replicates of each. for each IP i also have an input sample. I have used MACS2 callpeak to find enriched regions in both samples and would now like to find differentially enriched regions between the samples, but i am stuck with this memory error. I have tried adjusting the different parameters of both bgddiff and Diffbind (which are quite limited) to try and get a succesfull run, but it keeps returning the same error.

Maby it is simple impossible to run it on the public servers due to the memory issues, but it's a pretty standard analysis and i find it odd to have the tools listed on the public servers if they can't be used for anything.

Do anyone have any ideas as to how i can get this analysis to run on the public servers? any suggestions would be much appreciated.


error memory job choices chip-seq • 148 views
ADD COMMENTlink modified 10 weeks ago • written 10 weeks ago by johanvn10
gravatar for Jennifer Hillman Jackson
10 weeks ago by
United States
Jennifer Hillman Jackson25k wrote:


Jobs of your size are known to work correctly with these tools unless one of these issues are present:

  1. The datasets are very large.
  2. The custom reference genome used is highly fragmented.
  3. A formatting problem exists in one of the inputs.

If the data is simply very large, then options for using Galaxy are explained here:

Help with checking for common input problems that can result in job errors are described in the Support FAQs:

  2. Start here:

Galaxy tutorials, including those for ChIP-seq, that can be reviewed as a usage guide:

  2. Example:

Thanks! Jen, Galaxy team

ADD COMMENTlink written 10 weeks ago by Jennifer Hillman Jackson25k
gravatar for johanvn
10 weeks ago by
johanvn10 wrote:

Thanks Jen I still can't figure out what is wrong - my bedgraph files from MACS2 callpeak are between 4.5 and 5.5 GB in size. would you say these are to large to run bdgdiff on main?

ADD COMMENTlink written 10 weeks ago by johanvn10

Hello, I took a look at your history and didn't find any obvious format or data problems. The parameters used with MACS2 are simply creating results that are too large/complex to process with this tool with the resources available at the public server.

Some things to try:

Filter the MACS2 results down to a single chromosome and run the tool on that subset of the data. You also might want to review the QA/QC steps in the tutorial I linked above. Both will help you to learn more about your data quality, depth of sequencing, and test if tuning the MACS2 parameters more would help. The MACS2 manual and google group would both be good resources.

The alternative is to set up your own Galaxy. A Cloudman Galaxy might be a good choice, as you can scale up the memory when configuring your own server. Cloudman and other Galaxy options are linked above.

ADD REPLYlink written 10 weeks ago by Jennifer Hillman Jackson25k
Please log in to add an answer.


Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 109 users visited in the last hour