Smanga1974/Honours_GWAS_PRS_OSCC_SouthAfrica

Honours research project on GWAS and PRS of oral squamous cell carcinoma (OSCC) in a South African cohort. Analyses performed using the H3AGWAS pipeline, PLINK, GEMMA, PRSice-2, PRS-CSx, and visualized in R with ggplot2. # 🧬 Honours GWAS & PRS: OSCC in South Africa Welcome to the **Honours_GWAS_PRS_OSCC_SouthAfrica** repository! This project focuses on Genome-Wide Association Studies (GWAS) and Polygenic Risk Score (PRS) analysis for Oral Squamous Cell Carcinoma (OSCC) within South African populations. --- ## 🧑‍🔬 Project Overview This repository contains code, pipelines, and documentation for conducting GWAS and PRS analyses on OSCC datasets. The goal is to identify genetic variants associated with OSCC and develop risk prediction models tailored to South African cohorts. --- ## ✨ Features - **GWAS Analysis:** Quality control, association testing, and visualization of genome-wide data. - **PRS Pipeline:** Construction and evaluation of polygenic risk scores using state-of-the-art tools and methods. - **Advanced Data Science Workflows:** From data wrangling and statistical modeling to interpretation and reporting, leveraging best practices in data science and bioinformatics. - **Reproducible Pipelines:** Modular scripts and workflows for seamless, reproducible data analysis. - **Customizable & Scalable:** Designed for HPC clusters and adaptable for other complex disease studies and large-scale datasets. --- ## 🛠️ Tech Stack & Tools - **Languages:** R, Python, Bash - **Libraries:** ggplot2, data.table, pROC, qqman, pandas, scikit-learn, matplotlib - **Bioinformatics Tools:** PLINK, GEMMA, PRSice-2, PRScsx, H3AGWAS pipeline - **Other Tools:** Git, Linux, HPC clusters, Jupyter Notebooks --- ## 🚀 Data Science Highlights - **Statistical Genetics:** Applied advanced statistical methods (logistic regression, linear mixed models, multiple testing correction) to high-dimensional genomic data. Integrated polygenic risk scoring using clumping, thresholding, bayesian regression model, and ROC analysis to assess predictive performance and model disease risk. - **Data Visualization:** Generated publication-quality plots (Manhattan, QQ, PCA, ROC curves) for effective communication …