Statistical Analysis of Gene Expression

A comprehensive bioinformatics study focused on identifying differentially expressed genes (DEGs) in lung cancer datasets. This repository demonstrates professional-grade statistical analysis, data normalization, and multiple testing correction techniques.

Overview

This project performs an in-depth analysis of the GSE10072 dataset (Lung Tumor vs. Normal Lung). It follows a rigorous bioinformatics pipeline to ensure robust biological insights.

Key Features

• Exploratory Data Analysis (EDA): Descriptive statistics and visualization of gene expression distributions.
• Statistical Hypotheses: Implementation of two-sample t-tests to evaluate expression differences.
• Multiple Testing Correction: Application of the Benjamini-Hochberg (FDR) procedure to control for false discoveries in high-dimensional genomic data.
• Significant Gene Identification: Automated filtering and ranking of biologically relevant genes based on p-value and fold-change thresholds.

Methodology

1. Preprocessing: Handling series matrix files and log-transformation.
2. Analysis: Row-wise statistical computation using scipy and statsmodels.
3. Validation: Visualizing results to ensure statistical assumptions are met.

Contents

• statistical_tests.ipynb: The main production-ready Python notebook.
• GSE10072_series_matrix.txt: High-throughput gene expression data.

---

Developed as part of the Advanced Bioinformatics curriculum at Sharif University of Technology.