Lung Cancer Differential Gene Expression Analysis

Overview

This project performs a comprehensive differential gene expression (DEG) analysis on publicly available lung cancer microarray data, comparing gene expression profiles between tumour tissue and healthy lung tissue. The goal is to identify genes that are significantly upregulated or downregulated in lung cancer, providing molecular insights into tumourigenesis and potential therapeutic targets.

Lung cancer remains the leading cause of cancer-related mortality worldwide, accounting for approximately 1.8 million deaths annually. Understanding the transcriptomic landscape of lung tumours is critical for developing targeted diagnostics and therapies.

Dataset

Parameter	Details
GEO Accession	GSE19188
Platform	Affymetrix Human Genome U133 Plus 2.0 Array (GPL570)
Comparison	Tumour vs Healthy lung tissue
Source	NCBI Gene Expression Omnibus (GEO)

Methods

Raw GEO Data → Expression Matrix Extraction → Group Classification
→ Design Matrix → Contrast Definition → limma Linear Model
→ Empirical Bayes Statistics → FDR Correction → DEG Filtering
→ Gene Annotation → Visualisation → Results Export

Statistical Framework

• Package: limma (Linear Models for Microarray Analysis)
• Moderation: Empirical Bayes (eBayes)
• Multiple Testing Correction: Benjamini-Hochberg FDR
• Significance Thresholds: adj.P.Val < 0.05 AND |log2FC| > 1

Key Results

Metric	Value
Total Significant DEGs	3,672
Upregulated in Tumour	1,129
Downregulated in Tumour	2,543

Top Differentially Expressed Genes

Gene	Expression	Biological Relevance
TNXB	Differentially expressed	Extracellular matrix glycoprotein; influences tumour invasion and tissue remodelling
AGER	Differentially expressed	Receptor for advanced glycation end products (RAGE); linked to lung cancer progression and inflammation
ADAMTS8	Differentially expressed	Putative tumour suppressor; implicated in ECM remodelling in cancer
ADH1B	Differentially expressed	Alcohol dehydrogenase; reflects altered metabolic landscape of tumour cells
VEPH1	Differentially expressed	Emerging evidence links it to tumour regulatory pathways
LRRC36	Differentially expressed	Leucine-rich repeat protein involved in protein-protein interactions and signalling

Multiple probes mapping to the same gene (e.g., AGER, ADAMTS8) indicate consistent and robust expression changes across independent probe sets — strengthening confidence in these findings.

Visualisations

Volcano Plot

The volcano plot below visualises effect size (log2 fold change) against statistical significance (−log10 adjusted p-value). Genes in the upper corners represent the most biologically and statistically significant DEGs. The top 10 genes are annotated directly on the plot.

figures/volcano_plot.png

Heatmap

A heatmap of the top 10 most significant DEGs across all samples, with row scaling applied to normalise expression. Tumour samples are labelled in red, healthy samples in blue.

figures/heatmap_top10_genes.png

Repository Structure

Lung-Cancer-DEG-Analysis/
│
├── data/
│   └── GSE19188_series_matrix.txt.gz    # Raw GEO series matrix (download from NCBI)
│
├── scripts/
│   └── lung_cancer_analysis.R           # Full annotated R analysis pipeline
│
├── results/
│   ├── all_genes_results.csv            # All genes with full statistics
│   ├── significant_genes.csv            # Filtered significant DEGs
│   ├── upregulated_genes.csv            # Genes upregulated in tumour tissue
│   └── downregulated_genes.csv          # Genes downregulated in tumour tissue
│
└── figures/
    ├── volcano_plot.png                 # Volcano plot with top 10 annotated genes
    └── heatmap_top10_genes.png          # Heatmap of top 10 DEGs

How to Reproduce This Analysis

Prerequisites

Install the following R packages before running the analysis:

install.packages("BiocManager")
BiocManager::install(c("GEOquery", "limma", "hgu133plus2.db", "AnnotationDbi"))

Steps

1. Clone this repository
2. Download the GSE19188 dataset from NCBI GEO and place it in the data/ folder
3. Open scripts/lung_cancer_analysis.R in RStudio
4. Run the script end to end
5. Results will be saved to results/ and figures/

---

Biological Interpretation

The predominance of downregulated genes (2,543 vs 1,129 upregulated) is consistent with the widespread transcriptional silencing and tumour suppressor loss documented in lung cancer biology. Key findings include:

• AGER downregulation reflects disruption of normal pulmonary homeostasis and innate immune signalling in tumour tissue
• ADAMTS8 downregulation supports its role as a tumour suppressor restraining ECM remodelling in healthy lung tissue
• ADH1B dysregulation reflects the altered metabolic reprogramming characteristic of tumour cells (Warburg effect)
• Multiple probe sets mapping to the same genes strengthen confidence in these findings as genuine biological signals

Limitations

• Exploratory analysis — findings require functional validation
• No clinical metadata stratification (e.g., cancer subtype, stage, treatment history)
• Microarray technology offers less precise quantification than RNA-sequencing
• Future analyses should incorporate pathway enrichment (GO, KEGG) for deeper mechanistic insight

Author

Caroline Gachema

Data sourced from NCBI GEO — publicly available for research use.