search

rickhelmus / patRoon

Workflow solutions for mass-spectrometry based non-target analysis.
https://rickhelmus.github.io/patRoon/
GNU General Public License v3.0
61 stars 18 forks source link

Error: Finding features using XCMS #104

Closed LeonSaal closed 6 months ago

LeonSaal commented 7 months ago

Hi Rick,

I'm having a problem using XCMS for feature finding using the docker image (sha256:d1dfc1508d017c6fe870ae082d776b4d53e18eb3478323ec5368ef206d41b56c).

> fList <- findFeatures(anaInfo, "xcms3")
Verifying if your data is centroided... fstcore package v0.9.18
(OpenMP detected, using 16 threads)
Done!
Finding features with XCMS for 90 analyses ...
Loading raw data...
Error in reducer$value.cache[[as.character(idx)]] <- values : 
  wrong args for environment subassignment
In addition: Warning messages:
1: call dbDisconnect() when finished working with a connection 
2: In parallel::mccollect(wait = FALSE, timeout = 1) :
  1 parallel job did not deliver a result

I'm not sure what is happening here. Could be due to the huge number of files to be analysed. Do you have any idea on how to mitigate that? If it really is caused by the amount of computations, I thought about e.g. splitting the anaInfo prior to feature finding and using consensus() to merge the different featureGroups back together afterwards.

Kind regards,

Leon

rickhelmus commented 7 months ago

Hi Leon,

Not really sure either... but the error is definitely triggered outside patRoon. Running 90 analyses doesn't sound too extreme to , so I think it should be able to work. Perhaps you could try to disable parallelization for XCMS? Then you may get a more useful error message. Also, sometimes strange errors are triggered if an HRMS file is corrupt.

LeonSaal commented 7 months ago

Hi Rick,

thank you for the suggestions! I am trying it with parallel processing disabled.

Do you know of a way to check, whether the data is corrupt e.g. using tools from Proteowizard? I could not find any information on that.

Kind regards,

Leon

rickhelmus commented 6 months ago

Ah, sorry I missed your comment...

I am not really aware of any verification tools. And I also noticed that some files can be read by e.g. ProteoWizard but cannot by OpenMS and vice-versa, so this is a bit tricky. In the end I usually just try to process sub-batches of the files to determine which are OK or not.

LeonSaal commented 6 months ago

No worries, thanks! Without parallel processing it worked in the end, so I don't know what was wrong before, didn't seem to be corrupted files.