AI-Driven Rare Disease Genomics: Gene Discovery, Network Biology & Disease Gene Prioritization

Aim

Workshop Objectives

Workshop Structure

Workshop Structure

📅 Day 0: Pre-Workshop Research Setup & Dataset Preparation (30 Minutes)

Introduction to the selected rare disease, research objectives, and complete computational workflow.

One-to-one mentor guidance for dataset selection, downloading, and organization of genomic or transcriptomic data.

Exploring public databases, including NCBI GEO, ClinVar, Orphanet, and UniProt, for disease-related datasets.

Installation and configuration of Cytoscape, cytoHubba, and essential network-analysis plugins.

Google Colab and Python environment setup, including required libraries and prepared notebooks.

Installation guidance for AutoDock Vina and BIOVIA Discovery Studio Visualizer, including protein and ligand file preparation.

System compatibility checks, software troubleshooting, and verification of dataset accessibility.

🛠️ Hands-on Activity:

Mentor-Assisted Dataset Download & Computational Environment Setup

Participants will receive one-to-one mentor guidance to identify and download a suitable rare disease dataset, configure Google Colab and Cytoscape, and prepare the software required for gene expression analysis, PPI network construction, pathway enrichment, ADMET screening, and molecular docking. The session will ensure that participants have access to prepared datasets, computational notebooks, and required research tools before the workshop begins.

📅 Day 1: Rare Disease Genomics, Differential Gene Expression & Biomarker Discovery

Introduction to rare diseases, their genetic basis, molecular pathology, and applications in precision medicine.

Understanding genetic variants, disease-associated genes, and molecular mechanisms underlying rare genetic disorders.

Identification of rare disease-associated genes using OMIM, ClinVar, Orphanet, NCBI Gene, and UniProt.

Exploring publicly available transcriptomic datasets from NCBI Gene Expression Omnibus (GEO).

Gene expression data preprocessing, normalization, and disease-versus-control comparisons.

Identification of differentially expressed genes (DEGs) using statistical analysis and adjusted p-values.

Visualization of gene expression patterns using volcano plots, heatmaps, and expression distributions.

Integration of disease-associated genes and DEGs for candidate molecular biomarker identification.

Introduction to biomarker evaluation using ROC curves, AUC, and independent validation datasets.

🛠️ Hands-on Activity:

Rare Disease Gene Identification & Candidate Biomarker Discovery Using Google Colab

Participants will analyse a prepared rare disease gene expression dataset obtained from NCBI GEO using Google Colab. They will identify differentially expressed genes, generate volcano plots and heatmaps, and explore disease-associated expression patterns using Python, pandas, GEO2R/limma, matplotlib, and seaborn. The identified genes will be cross-referenced with ClinVar, Orphanet, and UniProt to shortlist candidate molecular biomarkers for further network and pathway analysis.

Practical Output: A curated disease-associated gene list, DEG results, volcano plot, gene expression heatmap, and candidate biomarker shortlist.

📅 Day 2: PPI Network Construction, Hub Gene Discovery & Disease Pathway Enrichment

Introduction to network medicine and protein–protein interaction (PPI) analysis in rare disease research.

Construction of disease-associated protein interaction networks using STRING.

Network visualization, protein interaction confidence scores, and biological interpretation using Cytoscape.

Identification of highly connected proteins and candidate hub genes using cytoHubba.

Network topology analysis using degree centrality, betweenness centrality, and Maximal Clique Centrality (MCC).

Identification of protein interaction modules and disease-associated clusters using MCODE.

Gene Ontology (GO) enrichment analysis for biological processes, molecular functions, and cellular components.

Identification of dysregulated molecular pathways using KEGG, Reactome, Enrichr, and g.

Integration of hub genes, enriched pathways, and disease associations for candidate biomarker and therapeutic target identification.

🛠️ Hands-on Activity:

PPI Network Construction, Hub Gene Identification & Pathway Analysis Using Cytoscape

Participants will use the candidate genes identified on Day 1 to construct a protein–protein interaction network using STRING and Cytoscape. They will identify hub genes using cytoHubba, analyse network modules using MCODE, and perform Gene Ontology, KEGG, and Reactome pathway enrichment analysis. The resulting molecular networks and enriched biological pathways will be interpreted to identify potential disease-associated biomarkers and therapeutic targets.

Practical Output: A visualized PPI network, ranked hub-gene list, network modules, enriched biological pathways, and shortlisted candidate therapeutic targets.

📅 Day 3: Pathway-Guided Drug Repurposing, ADMET Screening & Molecular Docking

Understanding the relationship between disease-associated hub genes, dysregulated molecular pathways, and therapeutic targets.

Identification of potential drug targets using disease–gene–pathway associations from Open Targets and ChEMBL.

Therapeutic target prioritization based on disease relevance, functional importance, and druggability.

Introduction to computational drug repurposing and identification of potential therapeutic compounds.

Retrieval of drug structures and molecular properties using PubChem and ChEMBL.

Drug-likeness and pharmacokinetic assessment using SwissADME and pkCSM.

Prediction of potential toxicity risks using ProTox 3.0.

Identification and retrieval of disease-associated protein structures using UniProt and RCSB PDB.

Protein and ligand preparation for structure-based molecular docking.

Molecular docking of shortlisted drug candidates using AutoDock Vina.

Visualization of protein–ligand binding interactions, hydrogen bonds, and docking poses using BIOVIA Discovery Studio Visualizer.

Integration of disease-pathway relevance, ADMET properties, toxicity profiles, and docking scores for candidate drug prioritization.

Introduction to AI-assisted drug screening using molecular descriptors, QSAR, and machine-learning models.

🛠️ Hands-on Activity:

Rare Disease Drug Candidate Screening, ADMET Analysis & Molecular Docking

Participants will select a disease-associated therapeutic target identified through the Day 2 pathway analysis and retrieve its protein structure using UniProt and RCSB PDB. Using Open Targets, ChEMBL, and PubChem, they will identify potential drug-repurposing candidates and evaluate their drug-likeness, ADMET properties, and predicted toxicity using SwissADME, pkCSM, and ProTox 3.0.

Participants will then perform molecular docking using AutoDock Vina with prepared protein and ligand files, visualize binding interactions using BIOVIA Discovery Studio Visualizer, and integrate pathway relevance, docking scores, and ADMET results in Google Colab to prioritize promising computational drug candidates for further experimental investigation.

Practical Output: A disease-associated therapeutic target shortlist, comparative ADMET and toxicity profiles, molecular docking results, protein–ligand interaction visualizations, and a ranked list of candidate repurposable drugs.

Who Should Enrol?

Important Dates

Workshop Outcomes