BiocParallel worker error in MAE QC step (create_matrix_dna_rna_cor.R) - could not find function "seqlevelsStyle"

[efe3506]

During the MonoallelicExpression (MAE) pipeline execution, the workflow fails at the MonoallelicExpression_pipeline_QC_create_matrix_dna_rna_cor_R rule. The error originates from a BiocParallel worker process failing to locate the seqlevelsStyle function, which belongs to the GenomeInfoDb package.

Error message:

RuleException:
CalledProcessError in file [path_to_snakefile], line [line_number]:
Command 'set -euo pipefail; Rscript --vanilla [path]/.snakemake/scripts/[script_hash].create_matrix_dna_rna_cor.R' returned non-zero exit status 1.

Error: BiocParallel errors
remote errors, element index: 1, 4, 7, 10, 13, 16, ...
unevaluated and other errors
first remote error:
Error in seqlevelsStyle(seqlevelsInUse(gr_dna)): could not find function "seqlevelsStyle"
Execution halted

Environment
​DROP Version: 1.6.1
​Conda Environment: Isolated conda environment (drop_1.6)
​OS: Ubuntu Server

​Expected behavior
The create_matrix_dna_rna_cor.R script should successfully generate the DNA/RNA correlation matrix. The required packages should be accessible to all parallel workers.
​Additional context and Troubleshooting
​The GenomeInfoDb package is properly installed and loads without any issues in the active drop_1.6 conda environment when tested interactively.
​Adding library(GenomeInfoDb) at the top of the create_matrix_dna_rna_cor.R script does not resolve the issue.
​This indicates a scope issue where the libraries are not being exported to or loaded within the BiocParallel worker processes. The library(GenomeInfoDb) call likely needs to be added directly inside the anonymous function passed to bplapply (or similar), or the environment needs to be properly exported to the workers.

[vyepez88]

Hi, can you double check you are using DROP 1.6.1?
We added that library call in that version and it is working for us. https://github.com/gagneurlab/drop/blob/master/drop/modules/mae-pipeline/QC/create_matrix_dna_rna_cor.R#L27

[efe3506]

Hi, thank you for your reply.

I checked my environment and drop --version indeed returns drop, version 1.6.1. I recently updated the package within this conda environment using the command conda install bioconda::drop.

Since my project directory was initialized, is it possible that updating the package via conda does not automatically overwrite or update the local Scripts/ folder in an existing project directory? If that is the case, is there a recommended command (such as drop update) to synchronize the existing project files with the updated 1.6.1 environment without breaking the current pipeline run?

Thanks for your help.

[vyepez88]

yes, there is drop update