Following lesson 2 April 2026 at 16:00 Professor Piotto will lead Module 1, Section 1.1
Following lesson 3 April 2026 at 16:00 Professor Carmine Ostacolo will lead Module 1, Section 1.3
Module 1 — Fundamentals of Drug Discovery
Core conceptual foundations for the pipeline: physicochemical intuition, thermodynamic interpretation, medicinal chemistry constraints, and pathway-level pharmacology.
1.1 Introduction to AI for Drug Discovery & Physical Chemistry for Drug Discovery
High-level orientation and physicochemical intuition.
9 h
1.1 Introduction to AI for Drug Discovery & Physical Chemistry for Drug Discovery
High-level orientation and physicochemical intuition.
Scope
High-level orientation and physicochemical intuition.
Description
This unit introduces the end-to-end drug discovery pipeline (target identification, hit discovery, lead optimization, preclinical development), highlighting where and why AI-based methods are impactful. The AI component is intentionally conceptual and non-algorithmic, focusing on use cases (virtual screening, QSAR, generative design, repurposing) rather than methods. The physical chemistry component provides an operational understanding of key molecular properties (pKa, logP/logD, solubility, polarity, lipophilicity) and their qualitative impact on permeability, exposure, and developability. Emphasis is placed on interpreting property trends rather than deriving equations.
1.2 Physical Chemistry: Thermodynamics and Molecular Interactions
Quantitative interpretation of molecular recognition.
6 h
1.2 Physical Chemistry: Thermodynamics and Molecular Interactions
Quantitative interpretation of molecular recognition.
Scope
Quantitative interpretation of molecular recognition.
Description
This unit formalizes molecular interactions using thermodynamic principles. Free energy of binding is decomposed into enthalpic and entropic contributions, with explicit discussion of solvent effects, conformational penalties, and enthalpy–entropy compensation. The goal is to enable students to interpret affinity changes observed in modeling or experiments, not to perform simulations.
1.3 Medicinal Chemistry: Drug-likeness and ADMET; Lipinski Rules and Molecular Descriptors
Multi-parameter optimization and chemical design rules.
18 h
1.3 Medicinal Chemistry: Drug-likeness and ADMET; Lipinski Rules and Molecular Descriptors
Multi-parameter optimization and chemical design rules.
Scope
Multi-parameter optimization and chemical design rules.
Description
This extended unit covers drug-likeness criteria across chemical space, including Rule-of-Five and beyond-Ro5 compounds. ADMET properties (absorption, metabolic stability, clearance, toxicity) are discussed in relation to structural features. Molecular descriptors and fingerprints are introduced strictly as chemical abstractions supporting SAR reasoning, not as ML features (covered later). Case studies illustrate trade-offs between potency and developability.
1.4 Pharmacology: Signaling Pathways
Biological context for target modulation.
12 h
1.4 Pharmacology: Signaling Pathways
Biological context for target modulation.
Scope
Biological context for target modulation.
Description
This unit introduces the pharmacokinetic and pharmacodynamic principles required to interpret drug response, including exposure, bioavailability, clearance, target engagement, and dose–response relationships. It also presents the main receptor classes and their associated signaling outputs, providing the biological and pharmacological basis for understanding how target modulation translates into efficacy and adverse effects.
Module 2 — AI for Drug Repurposing
From target hypotheses to mechanistic interpretation and translational strategy for repurposing.
2.1 Target Identification and Validation: Omics, CRISPR, Chemogenomics
Generation of target hypotheses.
9 h
2.1 Target Identification and Validation: Omics, CRISPR, Chemogenomics
Generation of target hypotheses.
Scope
Generation of target hypotheses.
Description
This unit focuses on identifying and prioritizing targets using omics data, functional genomics (CRISPR screens), and chemogenomic relationships. The emphasis is on interpreting experimental and computational evidence to justify target selection, rather than on downstream drug design.
2.2 Pharmacology: Receptor–Ligand Interactions and Drug Mechanisms
Mechanistic interpretation of drug action, including receptor and enzyme interactions..
9 h
2.2 Pharmacology: Receptor–Ligand Interactions and Drug Mechanisms
Mechanistic interpretation of drug action, including receptor and enzyme interactions..
Scope
Mechanistic interpretation of drug action.
Description
This unit introduces the fundamental principles of pharmacodynamics, including drug–receptor interactions, dose–response relationships, and agonism/antagonism. It explores receptor types, enzyme–ligand interactions, and the molecular forces governing binding. The course also covers basic signal transduction pathways, enzyme mechanisms and inhibition, and provides an introduction to structure–activity relationships (SAR) to understand how molecular structure influences biological activity.
2.3 Network Pharmacology: Gene–Gene Interactions
Systems-level integration of targets and pathways.
9 h
2.3 Network Pharmacology: Gene–Gene Interactions
Systems-level integration of targets and pathways.
Scope
Systems-level integration of targets and pathways.
Description
Here, biological systems are modeled as networks to contextualize drug effects beyond single targets. Concepts such as disease modules, network proximity, and pathway crosstalk are introduced. This unit explicitly builds on (but does not repeat) pathway biology from Module 1 and target identification from Module 2.1.
2.4 Computational and Translational Approaches to Drug Repurposing and Clinical Development
Translation from computational hypotheses to clinical strategy.
9 h
2.4 Computational and Translational Approaches to Drug Repurposing and Clinical Development
Translation from computational hypotheses to clinical strategy.
Scope
Translation from computational hypotheses to clinical strategy.
Description
This unit integrates molecular, biological, and clinical evidence to evaluate repurposing opportunities. Topics include real-world data, clinical trial design considerations, patient stratification, and regulatory constraints. The emphasis is on decision-making and risk assessment rather than algorithm development.
2.5 AI Applications in Structural Biology
From structural hypotheses to systems-level biomedical insight
6 h
2.5 AI Applications in Structural Biology
From structural hypotheses to systems-level biomedical insight
Scope
Translation from AI-based structural hypotheses to biological interpretation and research planning.
Description
This unit introduces AI-driven methods used to generate and interpret structural hypotheses in biomedical research. Topics include protein structure and complex prediction, binding-site characterization, docking and scoring with learned potentials, molecular representation learning (sequence/structure/graph), and AI-assisted analysis of cryo-EM, X-ray crystallography, and SAXS data. Students will also explore how structural biology interfaces with systems medicine through network-informed target prioritization and mechanism-of-action inference. Practical activities may include reproducing reference pipelines, critical evaluation of model uncertainty and bias, and defining research mini-projects suitable for internships or thesis work.
Module 3 — Statistics and Computational Methods
Statistical validity, classical ML evaluation, and foundational computational concepts for drug discovery data.
3.1 Biomedical Statistics and Data Preprocessing
Statistical validity and data integrity.
12 h
3.1 Biomedical Statistics and Data Preprocessing
Statistical validity and data integrity.
Scope
Statistical validity and data integrity.
Description
This unit covers statistical concepts essential for drug discovery datasets, including distributions, hypothesis testing, confidence intervals, and multiple testing. Data preprocessing focuses on avoiding methodological errors such as data leakage, improper splits, and biased sampling. This unit defines how data should be prepared, not how models are built.
3.2 Supervised and Unsupervised Machine Learning Methods
Classical ML modeling and evaluation.
18 h
3.2 Supervised and Unsupervised Machine Learning Methods
Classical ML modeling and evaluation.
Scope
Classical ML modeling and evaluation.
Description
This unit introduces supervised (regression, classification) and unsupervised (clustering, dimensionality reduction) ML methods commonly used in cheminformatics. Emphasis is placed on model selection, cross-validation, performance metrics, and error analysis. All evaluation methodology is confined to this unit to avoid repetition elsewhere.
3.3 Structural Bioinformatics
Structural data interpretation.
6 h
3.3 Structural Bioinformatics
Structural data interpretation.
Scope
Structural data interpretation.
Description
This unit focuses on protein and RNA structure quality, binding site identification, and structural variability. The objective is to enable critical assessment of structural inputs used later in modeling and simulation modules.
3.4 Data Integration and Ontologies
Semantic integration of heterogeneous data.
9 h
3.4 Data Integration and Ontologies
Semantic integration of heterogeneous data.
Scope
Semantic integration of heterogeneous data.
Description
This unit addresses how diverse biomedical datasets are harmonized using ontologies and standardized identifiers. Emphasis is on data consistency, provenance, and interpretability rather than numerical processing.
Module 4 — Molecular Modeling and Simulation
Structure-based workflows, dynamics, quantum/classical methods, and affinity/QSAR modeling.
4.1 Computational Modeling: Docking, Scoring, and Molecular Dynamics
End-to-end structure-based modeling workflows.
15 h
4.1 Computational Modeling: Docking, Scoring, and Molecular Dynamics
End-to-end structure-based modeling workflows.
Scope
End-to-end structure-based modeling workflows.
Description
This unit introduces core computational modeling approaches used in drug discovery, including molecular docking, scoring functions, and basic molecular dynamics simulations. The focus is on ligand–target interaction analysis, pose evaluation, and the interpretation of simulation results to support structure-based reasoning.
4.2 Advanced Molecular Dynamics
Sampling quality and uncertainty.
9 h
4.2 Advanced Molecular Dynamics
Sampling quality and uncertainty.
Scope
Sampling quality and uncertainty.
Description
This unit extends MD concepts to enhanced sampling methods, convergence analysis, and advanced trajectory interpretation. It is explicitly separated from basic MD to avoid repetition.
4.3 QM/MM Theory and Applications in RNA Systems
Quantum-level modeling of biomolecular processes.
12 h
4.3 QM/MM Theory and Applications in RNA Systems
Quantum-level modeling of biomolecular processes.
Scope
Quantum-level modeling of biomolecular processes.
Description
This unit introduces QM/MM theory with applications to RNA recognition and catalysis, highlighting when quantum effects are essential and how to interpret results critically.
4.4 Binding Affinity Prediction
Prediction and uncertainty of binding strength.
6 h
4.4 Binding Affinity Prediction
Prediction and uncertainty of binding strength.
Scope
Prediction and uncertainty of binding strength.
Description
This unit focuses on comparing physics-based and ML-based affinity prediction approaches, with emphasis on ranking reliability, error sources, and applicability domains.
4.5 QSAR Modeling
Statistical modeling of structure–activity relationships.
9 h
4.5 QSAR Modeling
Statistical modeling of structure–activity relationships.
Scope
Statistical modeling of structure–activity relationships.
Description
This unit presents QSAR workflows, including dataset curation, descriptor selection, model validation, and interpretability. QSAR is treated as a statistically grounded methodology distinct from modern deep learning.
Module 5 — Advanced AI Techniques
Neural representation learning, generative design, robust data engineering, and interpretability for regulated settings.
5.1 Deep Learning and Graph Neural Networks for Molecular Modeling
Representation learning for molecules.
12 h
5.1 Deep Learning and Graph Neural Networks for Molecular Modeling
Representation learning for molecules.
Scope
Representation learning for molecules.
Description
This unit introduces neural representations of molecules, focusing on graph-based models and their chemical interpretation. Evaluation concepts are assumed from Module 3 and not repeated.
5.2 Generative Models and Transformer-based Optimization for Molecular Design
De novo molecular generation and optimization.
12 h
5.2 Generative Models and Transformer-based Optimization for Molecular Design
De novo molecular generation and optimization.
Scope
De novo molecular generation and optimization.
Description
This unit covers generative AI methods for molecule design, emphasizing constraint handling, multi-objective optimization, and chemical validity.
5.3 Data Engineering for AI Pipelines in Drug Discovery
Reproducible and scalable AI workflows.
9 h
5.3 Data Engineering for AI Pipelines in Drug Discovery
Reproducible and scalable AI workflows.
Scope
Reproducible and scalable AI workflows.
Description
This unit addresses pipeline orchestration, data versioning, experiment tracking, and reproducibility, clearly separated from data preprocessing (Module 3.1).
5.4 Explainable AI for Clinical Processes and Drug Discovery
Interpretability and trust.
12 h
5.4 Explainable AI for Clinical Processes and Drug Discovery
Interpretability and trust.
Scope
Interpretability and trust.
Description
This unit focuses on explainability techniques, their limitations, and their role in regulatory and clinical contexts.
Module 6 — Industrial Applications and Case Studies
Industrial practice, governance, and commercialization topics complementing the technical curriculum.
6.1 Industrial Case Study: AI in Pharma
Real-world end-to-end industrial case studies.
9 h
6.1 Industrial Case Study: AI in Pharma
Real-world end-to-end industrial case studies.
Scope
Real-world end-to-end industrial case studies.
Description
Real-world end-to-end case studies illustrating integration of AI into pharmaceutical R&D.
6.2 IP Management and Innovation
Intellectual property strategy for AI-driven discovery.
6 h
6.2 IP Management and Innovation
Intellectual property strategy for AI-driven discovery.
Scope
Intellectual property strategy for AI-driven discovery.
Description
Foundations of intellectual property strategy in AI-driven drug discovery.
6.3 Data Security in AI-based Discovery
Governance and compliance for sensitive biomedical data.
6 h
6.3 Data Security in AI-based Discovery
Governance and compliance for sensitive biomedical data.
Scope
Governance and compliance for sensitive biomedical data.
Description
Governance, compliance, and secure handling of sensitive biomedical data.
6.4 Federated Learning and Privacy in Drug Discovery
Collaborative learning under privacy constraints.
9 h
6.4 Federated Learning and Privacy in Drug Discovery
Collaborative learning under privacy constraints.
Scope
Collaborative learning under privacy constraints.
Description
Collaborative learning under data privacy constraints, with pharmaceutical use cases.
6.5 AI Startups and Commercialization
Translation of AI technologies into ventures.
6 h
6.5 AI Startups and Commercialization
Translation of AI technologies into ventures.
Scope
Translation of AI technologies into ventures.
Description
Translation of AI technologies into sustainable biotech and pharma ventures.
Master's Program Professors
