Identify candidate motifs responsible of STR-initiating RNA

Introduction

This is a part of the study "Impact of transcription initiation at microsatellites on gene expression" M. Grapotte et al., 2024. In this part, we aim to identify candidate motifs responsible of STR-initiating RNA with HOMER findMotifs.pl.

Requirements

Nextflow need to be installed on your system. If it is not the case, you can install it with the following the instructions here.

Conda need to be installed on your system. If it is not the case, you can install it with the following the instructions here.

Download the data

The data used in this study is available here. The data is composed of 4 files and 1 directory:

• mnnModels/: a directory containing the hyperparameters and the parameters of the different MNN models.
• hg38.hipstr_reference.cage.500bp.around3end.fa: a fasta file containing the sequences of the 500bp around the 3' end of all STRs.
• hg38all_names_raw.npy : a numpy file containing the names of all STRs.
• hg38all_seqs_raw.npy : a numpy file containing the one-hot encoded sequences of all STRs.
• merged_results.txt : a TSV file listing all the STR-gene association with a correct prediction of the CAGE signal.

Those files need to be downloaded and placed in the data directory.

Launch the pipeline

To launch the pipeline, you need to execute the following command in the repository where the main.nf file is located:

nextflow main.nf

A local1p profile is used to run the pipeline on your local machine (option -profile local1p). It launches the pipeline with a single process. Some processes need a large amount of memory and can crash if you run the pipeline with too much parallel executions or on a machine with limited memory. The -resume option allows you to resume the pipeline from where it stopped if it was stopped for any reason.

Results

The results of the pipeline are located in the results directory. Pregenerated results are available here.

Issues

Use singularity instead of conda for IFB cluster

On the IFB cluster, there is a problem with the conda environment and the path to the python interpreter. To solve this problem, we need to use a singularity container. To do so, you need to execute the following command:

cd env/
singularity build GeneExprRegBySTR.simg Singularity
cd ..

Then, you can launch the pipeline normally with ifb profile (-profile ifb).

slurm account

Nextflow use the default slurm account to submit the jobs. If you want to use another account, you have to change the default account :

sacctmgr update user <your-login> set defaultaccount=<project-name>

out-of-memory while creating the conda environment

The pipeline can crash with the following error:

ERROR ~ Error executing process > 'GET_SEQ_NAMES_AND_ONE_HOT_BY_STR_CLASS (1)'

Caused by:
  Failed to create Conda environment
  command: mamba env create --prefix [DIR] --file [ENV_PATH]
  status : 137
  message:



 -- Check '.nextflow.log' file for details


WARN: Killing running tasks (1)

slurmstepd: error: Detected 1 oom-kill event(s) in StepId=37724626.0. Some of your processes may have been killed by the cgroup out-of-memory handler.
srun: error: cpu-node-3: task 0: Out Of Memory

Try to execute the command given after command: in the error message.

strict repo priority while creating the conda environment

When creating the conda environment, you may encounter the following error:

Encountered problems while solving:
  - package pytorch-2.2.0-py3.9_cuda12.1_cudnn8.9.2_0 is excluded by strict repo priority

It can be solved by executing the following command:

conda config --set channel_priority disabled

Then, try to create the environment again.

OMP: Error #15

On the IFB cluster, you may encounter the following error:

Command error:
  OMP: Error #15: Initializing libomp.so, but found unknown library already initialized.
  OMP: Hint This means that multiple copies of the OpenMP runtime have been linked into the program. That is dangerous, since it can degrade performance or cause incorrect results. The best thing to do is to ensure that only a single OpenMP runtime is linked into the process, e.g. by avoiding static linking of the OpenMP runtime in any library. As an unsafe, unsupported, undocumented workaround you can set the environment variable KMP_DUPLICATE_LIB_OK=TRUE to allow the program to continue to execute, but that may cause crashes or silently produce incorrect results. For more information, please see http://openmp.llvm.org/

Relaunching the pipeline with the resume option (-resume) seems to solve the problem.

citation

If you use this pipeline, please cite the following article:

@article{grapotte2024impact,
  title={Impact of transcription initiation at microsatellites on gene expression},
  author={Mathys Grapotte, Christophe Vroland, Diego Garrido-Martin, Alessio Vignoli, Lisa Calero, Quentin Bouvier, Mathilde Robin, Clément Chatelain, Laurent Bréhélin, Cédric Notredame, Roderic Guigo, Charles-Henri Lecellier},
  journal={...},
  year={2024},
  publisher={...}
}