Error for specific simulation replicate in `R_GET_GCTA_INTERVALS` process

[mckeowr1]

Error message from .nextflow.log

Caused by:
  Process `R_GET_GCTA_INTERVALS (5 22 0.8 gamma inbred pca cb.strain.set.290_bacteriophobic.pigeon_0.05)` terminated with an error exit status (140)


Command executed:

  Rscript --vanilla Get_GCTA_Intervals.R cb.strain.set.290_bacteriophobic.pigeon_0.05_Genotype_Matrix.tsv 5_22_0.8_0.05_gamma_cb.strain.set.290_bacteriophobic.pigeon_sims.pheno 5_22_0.8_0.05_gamma_cb.strain.set.290_bacteriophobic.pigeon_lmm-exact_inbred_pca.fastGWA \
      cb.strain.set.290_bacteriophobic.pigeon_0.05_total_independent_tests.txt 5 22 1000 150 0.8 BF cb.strain.set.290_bacteriophobic.pigeon 0.05 gamma inbred_pca
  
  cat <<-END_VERSIONS > versions.yml
  "R_GET_GCTA_INTERVALS":
      R: $( Rscript --version |& cut -f 4 )
  END_VERSIONS

Command exit status:
  140

Command output:
  [1] "100%"

Command error:
  
  Attaching package: ‘dplyr’
  
  The following objects are masked from ‘package:stats’:
  
      filter, lag
  
  The following objects are masked from ‘package:base’:
  
      intersect, setdiff, setequal, union
  
  Rows: 10380 Columns: 294
  ── Column specification ────────────────────────────────────────────────────────
  Delimiter: "\t"
  chr   (2): REF, ALT
  dbl (292): CHROM, POS, BRC20069, BRC20073, BRC20077, BRC20089, BRC20090, BRC...
  
  ℹ Use `spec()` to retrieve the full column specification for this data.
  ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
  Adding missing grouping variables: `trait`
  Warning message:
  In if (findPks$nBF == 1 & length(unique(findPks$CHROM)) == 1) { :
    the condition has length > 1 and only the first element will be used

Workdir: /scratch4/eande106/Ryan/2025_briggsae_gwas_strainset_sims/4f/3850e86048a4e8a969ab57cbca4ae0

[mckeowr1]

It appears that an error occurred within the R script for this specific simulation replicate. Almost all of the ~749 iterations of this process were successful, including other simulation replicates for this mapping panel.

[mckeowr1]

This actually looks like an issue with this process exceeding the runtime allotted. According to the NF report the duration of this process was 19 mins and 27 seconds and the allocated time for the process in 20m. I'm thinking this was killed by the scheduler.

This simulation replicate was clearly an outlier compared to all other iterations of this process. I'm thinking that it got hung up for some reason?

I'm not sure increasing the resouce allocation here is going to help resolve this issue.

@msauria what do you think here?

I'll dig into the actual processing script more to see if there is anything specific that might result in the abnormally long runtime.

[mckeowr1]

After making updates to increase the compute time and to ignore R_GET_GCTA_INTERVALS processes that failed in b53d594 and adding print statements for debugging in 3965d77 I reran the pipeline on RF.

All the R_GET_GCTA_INTERVALS processes were repeated, and this time, they were all completed successfully, including the specific simulation rep that had failed and prompted this issue.

I think this suggests that the error can be solved by adding a dynamic retry to increase the compute time for the job.

[mckeowr1]

Added dynamic runtime updates in e6de3cf

[mckeowr1]

To ensure this issue is fully resolved, I will run additional simulations with more mapping panels to check that this is no longer an error.