Polygenic Risk Score (PRS) Continuous Shrinkage (CS) Pipeline Using Snakemake

Overview

This Snakemake pipeline is designed for calculating polygenic risk scores (PRS) using PRS-CS Auto. It automates preprocessing of summary statistics, genotype data preparation, filtering with HapMap3 SNPs, scoring PRS, and removing the APOE region for Alzheimer's Disease research.

Workflow Steps

1. Munge Summary Statistics

   • Converts raw summary statistics into a format compatible with PRS-CS Auto.
   • Container: codewithrakshya/mungesumstats_docker

2. Process BIM/PVAR Files into PLINK Format

   • Generates BED/BIM/FAM files from BIM/PVAR files.
   • Uses plink.yaml conda environment

3. Prepare HapMap3 SNP List (Optional)

   • Prepares a list of HapMap3 SNPs for filtering purposes.

4. Filter Genotype Data

   • Filters genotype data to include only HapMap3 SNPs.
   • Uses PLINK for SNP extraction.

5. Run PRS-CS Auto

   • Executes PRS-CS Auto for each chromosome.
   • Produces PRS weights for subsequent analysis.

6. Concatenate Chromosome-Specific PRS Weights

   • Concatenates PRS weights from all chromosomes into a single file.

7. Remove APOE Region

   • Removes SNPs located in the APOE region (chr19:44,408,822-45,408,822).

8. Calculate PRS with PLINK

   • Scores individuals based on concatenated PRS weights using PLINK.

Prerequisites

• Snakemake v8.0+
• Singularity or Docker installed for containerized execution
• Conda environment set up for PLINK and PRS-CS (envs/plink.yaml, envs/prscs.yaml)

Configuration

Edit parameters in config/config.yaml:

• Cohort information
• Paths to summary statistics files
• Paths for reference genomes and other necessary input files

Execution

Run the pipeline using the command:

snakemake --use-conda --use-singularity --cores <num_cores>

Directory Structure

├── config
│   └── config.yaml
├── envs
│   ├── plink.yaml
│   └── prscs.yaml
├── scripts
│   └── munge_sumstats.R
├── resources
│   └── (input files)
├── results
│   └── prscs_results
└── Snakefile

Configuration (config.yaml)

Adjust paths, filenames, and parameters such as ref_dir, sample sizes, and chromosome lists directly within config.yaml.

Troubleshooting

• Ensure containers and conda environments are correctly set up.
• Verify paths in config.yaml and availability of summary statistics and genotype files.

For issues, review Snakemake logs in .snakemake/log/.

---

For questions or support, please contact Rakshya Sharma.