AI_and_Biotech_Project

Exploring RNA-Seq derived Biomarkers for colorectal cancer classification via Machine Learning.

---

👥 Project Members

• Yusuf Munir Aliyu
• Saniya Khurshid
• Fahad Sajjad
• Farha Tarique

---

📌 Project Overview

Colorectal cancer (CRC) is the third most prevalent malignancy worldwide and accounts for nearly 10% of all cancer-related deaths. Accurate early detection remains a challenge due to tumour heterogeneity and lack of reliable biomarkers.

This project integrates:

• RNA-seq preprocessing and normalisation
• Differential gene expression (DEG) analysis
• Feature selection (LASSO, ROC-AUC screening)
• Functional enrichment (GO, KEGG)
• Machine-learning classification (SVM)

The pipeline leverages both Galaxy (for raw FASTQ analysis) and R/Python for computational downstream analysis.

🔬 Hybrid Workflow Used:

✔ Galaxy → QC → trimming → alignment → quantification
✔ R + Python → DEG → biomarker selection → ML classification

---

🎯 Objectives

1. Identify differentially expressed genes (DEGs) between CRC tumour and normal tissues.

2. Screen biomarker genes using:

   • LASSO regression
   • ROC analysis
   • ML feature ranking

3. Build supervised ML models:

   • Support Vector Machine (SVM)

5. Provide a reproducible bioinformatics workflow for future CRC biomarker research.

---

🧬 Workflow

📁 Dataset Information

Accession ID: GSE156451
Samples: 144 total

• 72 colorectal cancer tumour tissues
• 72 matched adjacent normal tissues

Platform: Illumina RNA-seq
Data type: Raw FASTQ + processed gene counts

Raw dataset available from GEO:
🔗 https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE156451

---

📂 Repository Structure

CRC-Biomarker-Discovery/
│
├── 📁 Data/
│   ├── gene_counts_cleaned.csv
│   ├── metadata.csv
│   └── README.md
│
├── 📁 References/
│   └── Tool_References.md
│
├── 📁 Scripts/
│   ├── DESeq2_Normalization_and_Plotting.R
│   ├── GO_and_KEGG.R
│   ├── ML_script.ipynb
│   ├── biomarker_figures_script.Rmd
│   ├── to_merge_featurecounts.py
│   └── README.md
│
├── 📁 Results/
│   ├── 📁 DEG/
│   │   ├── volcano_plot.png
│   │   ├── heatmap_top50.png
│   │   └── DESeq2_results.csv
│   │
│   ├── 📁 Enrichment_Analysis/
│   │   ├── GO_BP_MF_CC.csv
│   │   ├── KEGG_pathways.csv
│   │   └── enrichment_plots/
│   │
│   ├── 📁 ML/
│   │   ├── LASSO_results.csv
│   │   ├── SVM_RFE_results.csv
│   │   ├── core_genes.csv
│   │   ├── ROC_curves.png
│   │   └── stability_scores.csv
│   │
│   ├── 📁 QC/
│   │   ├── fastqc_reports/
│   │   └── multiqc_report.html
│
├── 📁 Figures/
│   ├── workflow_diagram.png
│   ├── PCA_UMAP.png
│   └── biomarker_violin_density.png
│
├── 📁 Docs/
│   ├── Project_Overview.md
│   ├── Pipeline_Workflow.md
│   ├── QC_Guidelines.md
│   └── References.md
│
├── LICENSE
└── README.md

---

🧰 Tools Used

🧪 RNA-seq Processing (Galaxy platform)

• FastQC — Quality control
• Fastp — Trimming and filtering
• MultiQC — QC report summary
• Hisat2 — Genome alignment
• FeatureCounts — Gene-level quantification

📊 R Packages

• DESeq2 — Differential expression
• ClusterProfiler — GO/KEGG enrichment
• pROC — ROC curve analysis
• glmnet — LASSO regression

🤖 Python Packages

• numpy
• pandas
• scipy
• scikit-learn
• matplotlib / seaborn

---

📄 License This project is released under the MIT License. You are free to use, modify, and distribute with attribution.

📣 Citation Farha, T., Munir, Y. A., Saniya, K., Fahad, S. AI_and_Biotech_Project: ----

📬 Contact Email: tariquefarha@gmail.com