AI-Powered Lipid Nanoparticle Design for RNA Therapeutics & Gene Delivery

Aim

Workshop Objectives

Workshop Structure

📅 Day 1: LNP Design, Lipid Chemistry & RNA Delivery

Focus: Understand LNP architecture, lipid composition, formulation properties, and mechanisms of RNA delivery.

• LNP architecture and roles of ionizable lipids, phospholipids, cholesterol, and PEG-lipids
• Lipid physicochemical properties and structure–property–performance relationships
• RNA encapsulation, cellular uptake, endosomal escape, and intracellular release
• LNP applications in mRNA, siRNA, miRNA, and gene delivery
• Key formulation parameters: particle size, PDI, encapsulation efficiency, and delivery performance
• Introduction to computational and AI-guided LNP design

🛠️ Hands-on:

Analyse curated LNP datasets, generate molecular and formulation descriptors, compare LNP compositions, and prepare an ML-ready dataset.

Workflow:
Lipid Chemistry → LNP Formulation → RNA Encapsulation → Delivery Performance

🧰 Tools Covered:

PubChem, RDKit, Python, pandas, NumPy, Google Colab

📅 Day 2: Machine Learning for LNP Performance Prediction

Focus: Develop predictive models linking lipid composition and formulation features with LNP performance.

• LNP dataset curation, preprocessing, and feature engineering
• Molecular and formulation descriptors for predictive modelling
• Prediction of particle size, PDI, encapsulation efficiency, and delivery-related outcomes
• Random Forest and XGBoost-based predictive modelling
• Model training, validation, and cross-validation
• Performance evaluation using R², MAE, RMSE, and classification metrics
• Feature importance and identification of key formulation determinants

🛠️ Hands-on:

Build and validate an ML model for LNP performance prediction and identify formulation features influencing delivery.

Workflow:
LNP Dataset → Feature Engineering → ML Modelling → Validation → Feature Interpretation → Prediction

🧰 Tools Covered:

Python, RDKit, pandas, scikit-learn, Random Forest, XGBoost, Google Colab

📅 Day 3: AI-Guided LNP Optimisation & Candidate Prioritisation

Focus: Apply AI, explainable modelling, and multi-objective optimisation to prioritise promising LNP formulations.

• AI-guided optimisation of LNP composition and formulation parameters
• Prediction of performance for new and unseen formulations
• Explainable AI and SHAP-based interpretation of model predictions
• Virtual screening of LNP formulation libraries
• Multi-objective optimisation and candidate ranking
• Balancing formulation stability, encapsulation, and delivery performance
• RNA-specific delivery considerations and translational challenges
• Computational prediction, experimental validation, and model limitations

🛠️ Hands-on:

Screen virtual LNP formulations, interpret model predictions, optimise formulation parameters, and generate a prioritised candidate shortlist for experimental validation.

Workflow:
LNP Library → AI Prediction → Explainable AI → Optimisation → Candidate Ranking → Experimental Validation

🧰 Tools Covered:

Python, RDKit, XGBoost, SHAP, scikit-learn, pandas, Plotly, Google Colab

Who Should Enrol?

Important Dates

Workshop Outcomes