Structural Bioinformatics, AlphaFold & AI Drug Discovery
From 3D atomic coordinates to generative protein engineering. Master macromolecular visualization, AlphaFold2/3 Evoformer architectures, pLDDT/PAE confidence, ultra-fast ESMFold, virtual screening with AutoDock Vina, 100 ns GROMACS molecular dynamics, and de novo design with RFdiffusion.
Curriculum Modules
Click on any chapter below to explore complete code, stereochemical diagrams, and docking protocols.
1: Protein 3D Architecture, Coordinate Formats & PDB/mmCIF
Peptide bond geometry, Ramachandran dihedral angles, secondary structure motifs, PDB legacy format limitations, and mmCIF macromolecular data structures.
2: Molecular Visualization with PyMOL & ChimeraX
PyMOL CLI scripting, structural selections, ray tracing, cartoon/surface models, ligand contact pockets, and UCSF ChimeraX cryo-EM density mapping.
3: Classical Homology Modeling & Template Refinement
Comparative modeling principles, target-template alignments, SWISS-MODEL, MODELLER spatial restraint satisfaction, energy minimization, and PROCHECK validation.
4: The AlphaFold Revolution: Evoformer & Invariant Point Attention
AlphaFold2 deep learning architecture: MSA representation, Pair representation, triangular self-attention, and the 3D Structure Module with Invariant Point Attention (IPA).
5: Interpreting Structural Confidence: pLDDT, PAE & Disorder
Per-residue confidence metric (pLDDT bands), Predicted Aligned Error (PAE) matrices, domain orientations, and intrinsically disordered regions (IDRs).
6: Ultra-Fast Protein Folding: ESMFold, ColabFold & OmegaFold
Protein language models without MSAs (ESM-2 embeddings), ESMFold architecture, ColabFold MMseqs2 acceleration, and scaling to proteome-wide predictions.
7: Protein Complex Modeling & AlphaFold-Multimer
Quaternary structure prediction, AlphaFold-Multimer, cross-chain pairing, interface confidence (ipTM), predicting homooligomers and transient protein-protein interactions.
8: Small Molecule Cheminformatics & Chemical Databases
PubChem, ChEMBL, ZINC20, SMILES strings, InChIKey, RDKit Python cheminformatics library, Morgan fingerprints, Lipinski's Rule of 5, and ADMET filters.
9: Molecular Docking Principles, Receptor Prep & Grid Boxes
Rigid receptor vs flexible ligand, PROPKA protonation, Gasteiger charges, defining search grid bounding boxes, and PDBQT conversion.
10: Virtual Screening with AutoDock Vina & Smina
AutoDock Vina scoring function, parallel high-throughput screening, interpreting binding free energy (kcal/mol), and pose RMSD clustering.
11: Molecular Dynamics Simulations with GROMACS
Force fields (CHARMM/AMBER), periodic boundary solvent boxes, counter-ion neutralization, energy minimization, NVT/NPT equilibration, and production MD trajectories.
12: Generative De Novo Protein Design: RFdiffusion & ProteinMPNN
Denoising diffusion probabilistic models for protein backbones (RFdiffusion), motif scaffolding, de novo binder design, and sequence sequence design with ProteinMPNN.
13: Structure-Based Drug Design Case Study: Kinase Inhibitor Discovery
End-to-end case study: Kinase target preparation (EGFR / BRAF), pocket analysis, virtual screening of 100,000 compounds, binding pose refinement, and MD stability verification.