🧬 Epigenetic Analysis: DNA Methylation Patterns (Illumina 450K)

Identifying differentially methylated CpG sites between healthy and diseased individuals using Illumina HumanMethylation450k array data.

Course: DNA/RNA Dynamics
Platform: Illumina HumanMethylation450k Array
Language: R (Bioconductor)

---

📋 Project Overview

This project performs a full differential DNA methylation analysis pipeline, from raw IDAT files through quality control, normalization, statistical testing, and visualization. The goal is to detect CpG sites with significant methylation differences between healthy and diseased sample groups.

Analysis Pipeline

Raw IDAT Files
     ↓
Sample & Probe Quality Control
     ↓
Detection p-value Filtering (threshold: 0.05)
     ↓
Quantile Normalization (preprocessQuantile)
     ↓
Statistical Testing (t-tests per CpG)
     ↓
Correction for Multiple Testing
     ↓
Visualization & Reporting

---

📁 Repository Structure

epigenetic-methylation-450k/
├── analysis.R        # Complete methylation analysis pipeline
├── .gitignore        # Files to ignore (data, outputs)
└── README.md         # This file

Note: Raw IDAT files and sample sheets are not included due to file size. Place them in the working directory before running the script.

---

🚀 Getting Started

1. Clone the repository

git clone https://github.com/MahanBalooei/epigenetic-methylation-450k.git
cd epigenetic-methylation-450k

2. Install R dependencies

if (!require("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install(c(
  "minfi",
  "IlluminaHumanMethylation450kmanifest",
  "IlluminaHumanMethylation450kanno.ilmn12.hg19"
))

install.packages(c("gplots", "qqman"))

3. Prepare your data

Place your raw IDAT files (or methylation beta matrix) and sample sheet into the working directory. Update file paths and sample group labels in analysis.R as needed.

4. Run the analysis

Open analysis.R in RStudio and run the script, or from the terminal:

Rscript analysis.R

---

📦 Tools & Packages

Package	Purpose
minfi	Reading IDAT files, QC, normalization
IlluminaHumanMethylation450kmanifest	Array annotation
gplots	Heatmap visualization
qqman	Manhattan and QQ plots

---

📊 Visualizations Generated

The pipeline produces the following plots:

• Density plots — Beta and M value distributions (raw vs. normalized)
• 6-panel QC plot — Overall sample quality assessment
• PCA plots — Colored by group, sex, and Sentrix ID
• Boxplots — p-value distributions (raw, uncorrected, corrected)
• Volcano plot — Significant CpG sites by effect size and significance
• Manhattan plot — Genome-wide view of differential methylation
• Heatmap — Hierarchical clustering of top differentially methylated probes

---

🔬 Key Analysis Steps

Quality Control

• Per-sample and per-probe QC using minfi
• Detection p-value filtering at threshold 0.05 to remove unreliable probes

Normalization

• Quantile normalization via preprocessQuantile to reduce technical variation between arrays

Statistical Testing

• Pairwise t-tests for each CpG site between healthy and diseased groups
• Multiple testing correction applied to control false discovery rate

---

👥 Contributors

Name	GitHub
Elif Güler	@elif-guler
Eyip Sinay Dalmaz	@Dalmaz-ES
Simay Erol	@Simay9
Barkin Kemec	@mbkemec
Negin Nilforoosh	@neginnilforosh
Kimia Kanouni	@kanounik
Mahan Balooei	@MahanBalooei

---

📄 License

This project is licensed under the MIT License — see the LICENSE file for details.