elasticQTL

A reproducible QTL pipeline for PLINK-based association testing, LD clumping, nested elastic net modeling, stability filtering, and final variant annotation — plus an external scoring/refit workflow for WGS/WES/array cohorts.

---

Overview

elasticQTL provides two main workflows:

• Training: run association testing, LD clumping, genotype export, nested elastic net modeling, stability filtering, and final variant annotation
• Validation / external scoring: apply a trained model to external cohorts and optionally refit on the shared marker set before scoring

Use the entrypoints:

• pipeline for training
• validate for external scoring/refit

For full details, see:

• docs/PIPELINE.md — training workflow steps, nested CV, missing genotype policies, and outputs
• docs/VALIDATION.md — external cohort scoring, refit workflow, and cohort harmonization
• docs/INPUT_FORMATS.md — required inputs, config fields, and manifest formats
• docs/TROUBLESHOOTING.md — common errors and fixes

---

Quickstart

Managed install (Skyline / OpenOmics)

The pipeline is installed at:

/data/openomics/prod/elasticQTL/v0.1.0/

⚠️ Do not run on the head node. The pipeline will error immediately if attempted.

# Grab an interactive node
srun -N 1 -n 1 --time=8:00:00 --mem=64gb -c 4 --pty bash

# Add executables to PATH
export PATH="/data/openomics/prod/elasticQTL/v0.1.0/bin:${PATH}"

# View help menus
pipeline -h
validate -h

If run on the head node, you will see:

############################################################
ERROR: Running this pipeline on a head node is not allowed.
Please submit this as a job to the cluster or run it from
an interactive node. You can grab an interactive node with:

  srun -N 1 -n 1 --time=8:00:00 --mem=64gb -c 4 --pty bash
############################################################

---

Local clone

If you are setting up elasticQTL outside of a managed install, clone the repo directly:

# Clone
git clone https://github.com/your-org/elasticQTL.git
cd elasticQTL

# Make entrypoints executable
chmod +x bin/pipeline bin/validate

# Add to PATH (add this to your ~/.bashrc or ~/.bash_profile to persist)
export PATH="$(pwd)/bin:${PATH}"

Then verify the install:

pipeline -h
validate -h

---

Running the pipeline

# 1. Copy and edit a training config
cp config/study.template.env config/study.env
nano config/study.env

# 2. Run training
pipeline --config config/study.env

# 3. Copy and edit a validation config
cp config/validation.template.env config/validation.env
nano config/validation.env

# 4. Run external scoring/refit
validate --config config/validation.env

---

Contributing

PRs and issues are welcome. When filing a bug report, please include:

• The config used (redact paths if needed)
• The relevant logs/*.log files
• manifest/params_used.txt (Step 1) or the validation logs (Step 2)