Cheminformatics Course

Lesson 1 • Core Software Ecosystem

Surveys dominant open-source and commercial cheminformatics toolkits. Prepares students to select appropriate tools for specific tasks.

Lesson 2 • Chemical Data Types and Sources

Surveys molecular data formats and major public databases. Connects data literacy to downstream computational workflows.

Lesson 3 • What Is Cheminformatics

Defines cheminformatics and its scope within computational chemistry and biology. Positions the discipline relative to bioinformatics and data science.

Lesson 4 • Molecular Representation Basics

Introduces 1D, 2D, and 3D molecular representations. Provides the conceptual basis for all subsequent encoding and modelling tasks.

Lesson 1 • Format Interconversion and Validation

Teaches programmatic format conversion and error detection. Ensures data integrity when moving molecules between tools and pipelines.

Lesson 2 • SMILES and SMARTS in Depth

Covers advanced SMILES syntax and SMARTS pattern matching. Enables precise substructure queries and molecular filtering.

Lesson 3 • Large-Scale Molecular Collections

Covers efficient storage and retrieval of millions of compounds. Prepares students for virtual screening and database management tasks.

Lesson 4 • Structure File Formats

Examines MOL, SDF, PDB, CIF, and XYZ formats in detail. Connects format choice to application requirements in modelling and databases.

Lesson 1 • Physicochemical Property Descriptors

Covers logP, pKa, solubility, and related ADMET-relevant properties. Links computed properties to drug-likeness and bioavailability assessment.

Lesson 2 • Structural Fingerprints

Explains bit-vector and count fingerprints including ECFP, MACCS, and path-based types. Provides the foundation for similarity searching and machine learning.

Lesson 3 • Similarity Metrics and Searching

Applies Tanimoto, Dice, and cosine coefficients to fingerprint-based similarity. Enables nearest-neighbour searching in chemical space.

Lesson 4 • Constitutional and Topological Descriptors

Introduces atom counts, molecular weight, and graph-based topological indices. Establishes the simplest numerical encodings of molecular structure.

Lesson 1 • Chemical Cartridges and Extensions

Introduces chemistry-aware database extensions enabling substructure and similarity queries. Extends standard SQL with molecular search capabilities.

Lesson 2 • Compound Registration Systems

Explains deduplication, salt stripping, and standardisation in registration workflows. Ensures data integrity across large corporate or public compound collections.

Lesson 3 • Data Curation and Quality Control

Covers systematic detection and correction of errors in chemical datasets. Produces clean, analysis-ready compound libraries.

Lesson 4 • Public Database Integration

Demonstrates programmatic access to PubChem, ChEMBL, and ZINC via APIs. Enables automated data retrieval for virtual screening and SAR analysis.

Lesson 1 • QSAR Fundamentals and History

Traces QSAR from Hansch analysis to modern machine learning approaches. Establishes the theoretical basis for structure-activity modelling.

Lesson 2 • Classical QSAR Modelling Methods

Implements multiple linear regression, PLS, and ridge regression for QSAR. Provides interpretable baseline models for SAR exploration.

Lesson 3 • Dataset Preparation for QSAR

Addresses activity data curation, endpoint selection, and chemical space coverage. Directly impacts model reliability and applicability domain.

Lesson 4 • Machine Learning QSAR Models

Applies random forests, SVMs, and gradient boosting to QSAR datasets. Achieves higher predictive accuracy for complex structure-activity landscapes.

Lesson 1 • Running Docking Campaigns

Guides high-throughput docking of large compound libraries against prepared targets. Translates docking theory into practical screening execution.

Lesson 2 • Virtual Screening Concepts

Defines ligand-based and structure-based screening strategies and their trade-offs. Sets the strategic context for computational hit identification.

Lesson 3 • Consensus Scoring and Hit Selection

Combines multiple scoring methods to improve hit selection reliability. Reduces false positives before expensive experimental follow-up.

Lesson 4 • Molecular Docking Principles

Explains scoring functions, search algorithms, and binding site preparation. Provides the mechanistic understanding needed to run docking correctly.

Lesson 1 • Trajectory Analysis

Analyses RMSD, RMSF, hydrogen bonds, and contact maps from trajectories. Extracts structural and dynamic insights relevant to binding.

Lesson 2 • Binding Free Energy Calculations

Applies MM-PBSA, MM-GBSA, and FEP methods to estimate binding affinities. Connects simulation data to quantitative affinity predictions.

Lesson 3 • MD Simulation Fundamentals

Introduces force fields, equations of motion, and ensemble types. Establishes the physical basis for interpreting simulation trajectories.

Lesson 4 • System Preparation for MD

Covers protein and ligand parameterisation, solvation, and ionisation. Correct preparation is prerequisite to meaningful simulation results.

Lesson 1 • Graph Neural Networks for Properties

Implements GCN, GAT, and MPNN architectures for molecular property prediction. Achieves state-of-the-art accuracy on benchmark datasets.

Lesson 2 • Generative Models for De Novo Design

Trains VAEs, GANs, and flow-based models to generate novel drug-like molecules. Enables goal-directed molecular generation with property constraints.

Lesson 3 • Model Interpretability and Deployment

Applies GradCAM, attention weights, and SHAP to explain deep learning predictions. Prepares models for integration into screening and design pipelines.

Lesson 4 • Molecular Graph Representations

Encodes molecules as graphs with atom nodes and bond edges for neural networks. Provides the data structure foundation for all graph-based models.