HAPNEST GWAS – Ancestry-Specific Genome-Wide Association Study

This repository contains all relevant code, documentation, and results for our project in the course Biomedical Data Types at Hasso Plattner Institute.

The project focuses on performing a genome-wide association study (GWAS) using the synthetic HAPNEST dataset. We analyze genetic associations separately for two ancestry groups – European (EUR) and African (AFR) – and combine the results through meta-analysis. The goal is to explore ancestry-specific patterns in genetic data and assess how population structure impacts GWAS outcomes.

---

Repository Structure

HAPNEST_GWAS/
├── data/
│   ├── raw/                # Original input data (PLINK files per chromosome)
│   ├── processed/          # Filtered & QC'd data per ancestry group
│   ├── maps/               # rsID mapping files per chromosome
│   └── results/            # results and visualizations
│
├── notebooks/
│   ├── 01_preprocessing.ipynb          # Data filtering, phenotype assignment and quality control
│   ├── 02_gwas_analysis.ipynb          # GWAS per ancestry and meta-analysis
│   └── 03_results_visualization.ipynb  # Manhattan plots and summary visuals
│
├── requirements.txt                    # Python dependencies
└── README.md

How to Run

1. Clone this repository and set up a python virtual environment.
2. Install dependencies: pip install -r requirements.txt.
3. Download PLINK into the project folder.
4. Download the HAPNEST dataset manually into data/raw/.
5. Run the preprocessing, QC and GWAS for both ancestries AFR and EUR.
6. Run the meta analysis.
7. Create plots & visualizations with 03_results_visualization.ipynb.

Tools Used

• PLINK v1.9 – for QC, GWAS, and meta-analysis (Slifer, 2018)
• Python 3.11 – data handling and preprocessing
• Libraries listed in requirements.txt
• Jupyter Notebook – reproducible workflow and step-by-step analysis

License & Credits

This project was developed as part of the course Biomedical Data Types at HPI. Dataset credit: HAPNEST project (S-BSST936, EBI). Code and content are for academic use only.

Authors

• Maximilian Kalff
• Emil Kobel