[Bug]: 3a halts at calculate DunedinPACE

[ks164]

Contact Details

ks164@duke.edu

Scripts

03a-methylation_variables.sh

What happened?

script runs successfully until calculating DunedinPACE:

Predicting, adjusting and standardizing DunedinPACE Loading required namespace: preprocessCore Error in PredAgeDun$DunedinPACE : $ operator is invalid for atomic vectors Calls: main -> data.frame Execution halted

How can the bug be reproduced?

Don't know if this is the relevant code, but bug not reproduced running DunedinPACE calculation outside of ./03a-methylation_variables.sh

i..e the following code from dnamage.R script directly on norm.beta (from methylation_no_outliers.Robj) generates paceresult dataframe:

message("Predicting, adjusting and standardizing DunedinPACE")######################################## PredAgeDun <- try(PACEProjector(norm.beta, proportionOfProbesRequired=0.7), silent = TRUE) paceresult <- data.frame("IID" = names(PredAgeDun$DunedinPACE), "PredAge" = PredAgeDun$DunedinPACE)

and

is.atomic(PredAgeDun) [1] FALSE

R version

4.3.3 (February, 2024)

Python version

None

Relevant log output

No response

[SiyiSEA]

Hi,

Thank you for running the pipeline.

The error has been fixed in the branch of 3a-halts-at-calculate-DunedinPACE and I've create a pull request for this issue.

Please be patient for merging and you could run the script again.

Best, Siyi

[SiyiSEA]

Hi @ks164 ,

The main repo has been updated.

Please run the git pull before you run the pipeline.

Thank you.

Best, Siyi

[ks164]

Thank you so much Siyi!

Unfortunately, the script still fails, this time with the error:

Error in PredAgeDun[["DunedinPACE"]] : subscript out of bounds Calls: main -> data.frame Execution halted

Could it be that the PACEprojector is returning an empty list?

[SiyiSEA]

Hi,

That's fine. Thanks for cooperating with me.

Could you please run print(PredAgeDun) in your R terminal? If there's an error message, it might be due to an issue with the PACEprojector function.

Since the PACEprojector belongs to the DunedinPACE package, could you also check if the DunedinPACE package has been installed successfully?

If not, please reinstall it by following the instructions on the wiki (https://github.com/genetics-of-dna-methylation-consortium/godmc_phase2/wiki/System-requirements). If it is installed, could you please check the version of the DunedinPACE package by typing the following command in your R terminal:

# check the package version
packageVersion('DunedinPACE')

Please don't hesitate to contact me.

Best, Siyi

[ks164]

Thanks again!

I've caught the issue running through the code a line at a time: it's a problem with preprocessCore

Error in preprocessCore::normalize.quantiles.use.target(betas.mat, target = mPACE_Models$gold_standard_means[[model_name]]) : ERROR; return code from pthread_create() is 22

e.g. https://github.com/Bioconductor/bioconductor_docker/issues/22

I use conda environments and I think it might only be an issue in that instance. I will try to manually reinstall and downgrade openblas

[SiyiSEA]

Hi,

That's cool.

There is another solution as mentioned in the wiki. I wish this could help you.

I'll enhance my code to make sure the error message will show up. Thank you a lot.

Best, Siyi

[ks164]

Thanks!

the threading.ok() came back as TRUE for me, so I think that's not the problem in this case.

I think it's an issue with conda environments- I recall having to modify the openblas/libopenblas library one other time to get preprocessCore to work. Other times it's fine...

[ks164]

The bug is fixed by manual install of preprocessCore from Github with threading disabled

  - name: Manually install preprocessCore
    run: |
      git clone https://github.com/bmbolstad/preprocessCore.git
      cd preprocessCore
      R CMD INSTALL --configure-args="--disable-threading"  .

I have now successfully run module 03a

[SiyiSEA]

Well done! I will update the wiki about the preprocessCore.