Dockerfile.gwama

scripts/run_gwama/Dockerfile.gwama

@@ -0,0 +1,39 @@
+FROM ubuntu:24.04
+
+# install system dependencies
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends \
+    python3>=2.2.0 \
+    python3-pandas>=1.26.0 \
+    python3-scipy>=1.12.0 \
+    wget \
+    unzip \
+    zlib1g-dev \
+    build-essential \
+    && \
+    apt-get clean \
+    && \
+    rm -rf /var/lib/apt/lists/*
+
+# install GWAMA
+WORKDIR /tmp
+RUN wget --no-check-certificate https://www.geenivaramu.ee/tools/GWAMA_v2.2.2.zip && \
+    unzip -d GWAMA GWAMA_v2.2.2.zip && \
+    cd GWAMA && \
+    make && \
+    chmod +x GWAMA && \
+    mv GWAMA /usr/local/bin/ && \
+    cd .. && \
+    rm -rf GWAMA GWAMA_v2.2.2.zip
+
+# copy conversion script and entrypoint
+WORKDIR /tools
+COPY regenie_to_gwama.py .
+COPY entrypoint.gwama.sh .
+RUN chmod +x regenie_to_gwama.py entrypoint.gwama.sh
+
+# Set entrypoint
+ENTRYPOINT ["/tools/entrypoint.gwama.sh"]
+CMD ["--help"]
+
+WORKDIR /home

scripts/run_gwama/README.md

@@ -1,10 +1,138 @@
-# Run GWAMA meta-analysis
+# GWAMA Meta-Analysis Docker Container
 
-## Download and build GWAMA
+This container automates GWAS meta-analysis using GWAMA, converting REGENIE output to the required GWAMA format and performing meta-analysis across multiple sites.
 
-The `GWAMA` binary need to exist on the server site where the meta-analysis is run.
+## Quick Start
+
+### Prerequisites
+
+- Docker installed
+- REGENIE output files (.regenie format) from each site
+
+
+### Basic Usage
+
+```bash
+docker run --platform=linux/amd64 \
+  -v /path/to/data:/data \
+  -v /path/to/output:/out \
+  ghcr.io/collaborativebioinformatics/gwama \
+  -o /out <mode> <output_prefix> <site1_regenie> [site2_regenie] ...
+```
+
+**Example usage:** (using binary trait GWAS in this repository)
 
 ```{bash}
+# build the Docker image locally
+cd scripts/run_gwama
+docker build --platform=linux/amd64 -t ghcr.io/collaborativebioinformatics/gwama -f Dockerfile.gwama .
+
+# run (help function)
+docker run --platform=linux/amd64 \
+    -it ghcr.io/collaborativebioinformatics/gwama --help
+...
+
+# run (example) - should put output in working directory 
+docker run --platform=linux/amd64 \
+    -v <path/to/FedGen>/resources/site1_gwas_results:/data \
+    -v <path/to/FedGen>/scripts/run_gwama/:/out \
+    -it ghcr.io/collaborativebioinformatics/gwama -o /out or gwama_meta /data/regenie_step2_Phen1.regenie
+```
+
+
+**Arguments:**
+- `mode`: `or` for binary traits or `qt` for quantitative traits
+- `output_prefix`: Prefix for output files
+- `regenie_files`: Paths to REGENIE output files (use absolute paths)
+
+### Example: Binary Trait (Case-Control)
+
+```bash
+docker run --platform=linux/amd64 \
+  -v /Users/espehage/resources:/data \
+  -v /Users/espehage/output:/out \
+  ghcr.io/collaborativebioinformatics/gwama \
+  -o /out or meta_analysis /data/site1.regenie /data/site2.regenie /data/site3.regenie
+```
+
+### Example: Quantitative Trait
+
+```bash
+docker run --platform=linux/amd64 \
+  -v /Users/espehage/resources:/data \
+  -v /Users/espehage/output:/out \
+  ghcr.io/collaborativebioinformatics/gwama \
+  -o /out qt meta_results /data/site1.regenie /data/site2.regenie
+```
+
+## Important: Volume Mounting
+
+The container **cannot access relative paths** from the host. You must:
+
+1. Use **absolute paths** to data files
+2. Mount directories with `-v /host/absolute/path:/container/path`
+3. Reference files by their mounted paths inside the container
+
+**❌ This will fail (relative paths don't work in containers):**
+```bash
+docker run ... gwama or meta ../../resources/site1.regenie
+```
+
+**✅ This works (absolute path or mounted volume):**
+```bash
+docker run -v /Users/espehage/resources:/data ... gwama or meta /data/site1.regenie
+```
+
+## Output Files
+
+Results are saved in the container's `/home/` directory by default, or in the directory specified with `-o/--outdir`.
+
+- `<output_prefix>.out` - Main GWAMA meta-analysis results
+- `<output_prefix>.err.out` - Error log from GWAMA
+- `<output_prefix>.in` - GWAMA input file list
+- `<output_prefix>_site*.txt` - Converted GWAMA format files per site
+
+To access results from your host, mount an output volume:
+
+```bash
+docker run --platform=linux/amd64 \
+  -v /data/gwas:/input \
+  -v /data/output:/out \
+  ghcr.io/collaborativebioinformatics/gwama \
+  -o /out or meta /input/site1.regenie /input/site2.regenie
+```
+
+Then retrieve results from `/output/`.
+
+## Help
+
+View the help message:
+
+```bash
+docker run ghcr.io/collaborativebioinformatics/gwama --help
+```
+
+## Building Locally
+
+```bash
+cd scripts/run_gwama
+docker build --platform=linux/amd64 -t gwama:local -f Dockerfile.gwama .
+docker run gwama:local or meta /data/site1.regenie
+```
+
+## Pipeline Steps
+
+The entrypoint automatically:
+
+1. **Converts** each site's REGENIE output to GWAMA format using `regenie_to_gwama.py`
+2. **Creates** GWAMA input file list
+3. **Runs** GWAMA meta-analysis with appropriate parameters for the trait type
+
+## Manual Setup (if not using Docker)
+
+### Download and build GWAMA
+
+```bash
 wget https://www.geenivaramu.ee/tools/GWAMA_v2.2.2.zip
 unzip -d GWAMA GWAMA_v2.2.2.zip
 cd GWAMA
@@ -13,26 +141,22 @@ chmod +x GWAMA
 cd ..
 ```
 
-## Run GWAMA meta-analysis (Regenie data format)
-
 ### Convert Regenie output to GWAMA input format
 
-```{bash}
+```bash
 export SITE=1
 export DATA_PATH="../../resources/site${SITE}_gwas_results"
 export FILEPREFIX="regenie_step2_Phen1.regenie"
 
-python3 regenie_to_gwama.py  \
+python3 regenie_to_gwama.py \
     "${DATA_PATH}/${FILEPREFIX}" \
     "site${SITE}_for_gwama.txt" \
     "or"
 ```
 
-### create input file list for GWAMA
-
-Should contain data from all 10 sites, here shown for site 1 only for brevity.
+### Create input file list for GWAMA
 
-```{bash}
+```bash
 echo site1_for_gwama.txt > gwama.in
 ```
 

scripts/run_gwama/entrypoint.gwama.sh

@@ -0,0 +1,172 @@
+#!/bin/bash
+
+set -euo pipefail
+
+# GWAMA meta-analysis entrypoint
+# Converts REGENIE results to GWAMA format and runs meta-analysis
+#
+# Usage:
+#   entrypoint.gwama.sh [--outdir DIR|-o DIR] <mode> <output_prefix> <site1_regenie_file> [site2_regenie_file] ...
+#
+# Arguments:
+#   mode: 'or' for odds ratio (case-control) or 'qt' for quantitative trait
+#   output_prefix: prefix for GWAMA output files
+#   regenie_files: paths to .regenie output files (one per site)
+#
+# Example:
+#   entrypoint.gwama.sh -o /output or gwama_results /data/site1.regenie /data/site2.regenie
+
+# Show help
+if [ $# -eq 0 ] || [[ "$1" == "--help" || "$1" == "-h" ]]; then
+    cat << 'EOF'
+GWAMA Meta-Analysis Pipeline
+
+Usage:
+  entrypoint.gwama.sh [--outdir DIR|-o DIR] <mode> <output_prefix> <site1_regenie_file> [site2_regenie_file] ...
+
+Arguments:
+  mode: 'or' for odds ratio (case-control) or 'qt' for quantitative trait
+  output_prefix: prefix for GWAMA output files
+  regenie_files: paths to .regenie output files (one per site, use absolute paths or mount volumes)
+  --outdir, -o: output directory inside container (default: /home)
+
+Example:
+  docker run -v /path/to/data:/data -v /path/to/output:/out ghcr.io/collaborativebioinformatics/gwama \
+    -o /out or gwama_results /data/site1.regenie /data/site2.regenie
+
+Note:
+  Use absolute paths or mount input data with -v flag. Relative paths from host do not work in containers.
+EOF
+    exit 0
+fi
+
+# Defaults
+OUTDIR="/home"
+
+# Collect positional args after consuming options (options can appear anywhere)
+POSITIONAL=()
+while [[ $# -gt 0 ]]; do
+    case "$1" in
+        -o|--outdir)
+            if [[ $# -lt 2 ]]; then
+                echo "Error: --outdir requires a directory path"
+                exit 1
+            fi
+            OUTDIR="$2"
+            shift 2
+            ;;
+        -h|--help)
+            exec "$0" --help
+            ;;
+        --)
+            shift
+            while [[ $# -gt 0 ]]; do
+                POSITIONAL+=("$1")
+                shift
+            done
+            ;;
+        -*)
+            echo "Error: Unknown option: $1"
+            echo "Run with --help for usage."
+            exit 1
+            ;;
+        *)
+            POSITIONAL+=("$1")
+            shift
+            ;;
+    esac
+done
+
+if [ ${#POSITIONAL[@]} -lt 3 ]; then
+    echo "Error: Insufficient arguments"
+    echo "Usage: $0 [--outdir DIR|-o DIR] <mode> <output_prefix> <site1_regenie_file> [site2_regenie_file] ..."
+    echo "Run with --help for more information"
+    exit 1
+fi
+
+mode="${POSITIONAL[0]}"
+output_prefix="${POSITIONAL[1]}"
+regenie_files=("${POSITIONAL[@]:2}")
+
+# Validate mode
+if [[ ! "$mode" =~ ^(or|qt)$ ]]; then
+    echo "Error: mode must be 'or' or 'qt', got '$mode'"
+    exit 1
+fi
+
+echo "======================================"
+echo "GWAMA Meta-Analysis Pipeline"
+echo "======================================"
+echo "Mode: $mode"
+echo "Output prefix: $output_prefix"
+echo "Number of sites: ${#regenie_files[@]}"
+echo ""
+
+# Create/ensure output directory and work there
+WORKDIR="$OUTDIR"
+mkdir -p "$WORKDIR"
+cd "$WORKDIR"
+
+# Convert each site's REGENIE output to GWAMA format
+echo "Step 1: Converting REGENIE outputs to GWAMA format..."
+gwama_input_files=()
+
+for i in "${!regenie_files[@]}"; do
+    site_num=$((i + 1))
+    input_file="${regenie_files[$i]}"
+    output_file="${output_prefix}_site${site_num}.txt"
+
+    if [ ! -f "$input_file" ]; then
+        echo "Error: Input file not found: $input_file"
+        echo ""
+        echo "Troubleshooting:"
+        echo "  - Ensure file exists at the specified path"
+        echo "  - Use absolute paths inside the container"
+        echo "  - Mount data directory with: docker run -v /host/path:/data ..."
+        echo "  - Then reference files as: /data/filename"
+        exit 1
+    fi
+
+    echo "  Site $site_num: Converting $input_file..."
+    python3 /tools/regenie_to_gwama.py "$input_file" "$output_file" "$mode"
+    gwama_input_files+=("$output_file")
+done
+
+echo ""
+echo "Step 2: Creating GWAMA input file list..."
+gwama_input_list="${output_prefix}.in"
+for file in "${gwama_input_files[@]}"; do
+    echo "$file" >> "$gwama_input_list"
+done
+echo "Created: $gwama_input_list"
+echo ""
+
+echo "Step 3: Running GWAMA meta-analysis..."
+if [ "$mode" == "or" ]; then
+    GWAMA \
+        -i "$gwama_input_list" \
+        --output "$output_prefix" \
+        --name_marker MARKERNAME \
+        --name_ea EA \
+        --name_nea NEA \
+        --name_or OR \
+        --name_or_95l OR_95L \
+        --name_or_95u OR_95U
+else
+    # Quantitative trait mode (no OR fields)
+    GWAMA \
+        -i "$gwama_input_list" \
+        --output "$output_prefix" \
+        --name_marker MARKERNAME \
+        --name_ea EA \
+        --name_nea NEA
+fi
+
+echo ""
+echo "======================================"
+echo "GWAMA meta-analysis completed!"
+echo "======================================"
+echo "Results saved to: $WORKDIR"
+echo "  - Main output: ${output_prefix}.out"
+echo "  - Error log: ${output_prefix}.err.out"
+echo ""

scripts/run_gwama/regenie_to_gwama.py

@@ -38,7 +38,7 @@ def parse_args():
 def main():
     args = parse_args()
 
-    gwas = pd.read_csv(args.input, sep="\s+")
+    gwas = pd.read_csv(args.input, sep="\\s+")
     if args.mode == 'or':
         z_score = norm.ppf(0.975) # 97.5 percentile
         gwas['OR'] = np.exp(gwas['BETA'])