AI drug Development foundation
Pharmaceutical AI Drug discovery
Full course curriculum
Goal: Go from Zero to building AI models that design new drugs in
12-14 days.
Foundations - Week 1-2
Biology +
Chemistry +
Python for AI*
Biology:
Cell,
proteins,
genes,
disease pathways,
drug targets
Chemistry:
SMILES,
molecular graphs,
functional groups,
Lipinski's Rule of 5
Python:
NumPy,
Pandas,
Matplotlib,
RDKit basics
ML Basics:
Regression,
Classification,
Train/Test split,
Overfitting
Module 2: Cheminformatics - Week 3
How computers "see" molecules
Molecular representations:
SMILES,
InChI,
Graphs,
Fingerprints ECFP
Molecular descriptors:
MW,
logP,
TPSA
Databases: ChEMBL, PubChem, ZINC,
DrugBank, PDB
Tools: RDKit, OpenBabel
Lab: Download 10,000 molecules from
ChEMBL
Module 3: AI for Target Discovery - Week 4
Find the right protein to drug
Omics data: Genomics,
Transcriptomics, Proteomics
Protein structure prediction: AlphaFold2, ColabFold, RoseTTAFold
Protein-ligand binding sites
Network pharmacology + PPI networks
Lab: Predict structure of COVID-19 main
protease
Module 4: Virtual Screening with AI - Week 5-6*
Find "hits" that bind to target
Molecular Docking:
AutoDock Vina,
Glide
AI Scoring Functions:
DeepDock,
EquiBind,
DiffDock
Ligand-based VS: Similarity
search
Graph Neural Networks GNNs for molecules
Lab: Screen 1M compounds against 1
target
Module 5: Generative AI for De Novo Design -
Week 7-8*
Create brand new molecules
VAEs, GANs, Diffusion Models
for molecules
Transformers:
ChemBERTa,
MolGPT
Reinforcement Learning:
REINVENT,
GCPN
Scaffold hopping + Fragment linking
Lab: Generate 1000 novel inhibitors for
target
Module 6: ADMET Prediction - Week 9
Predict if drug will fail in body
ADME: Absorption,
Distribution,
Metabolism,
Excretion,
Toxicity
Predict:
Solubility,
hERG,
Liver toxicity,
BBB,
CYP inhibitiona
Datasets:
Tox21,
ToxCast,
ADMETlab
Tools:
DeepChem,
Chemprop,
ADMET-AI
Lab: Train model to predict toxicity
Module 7: Clinical AI + Drug Repurposing - Week
10
Patient stratification with ML
Clinical trial optimization
Knowledge Graphs: Hetionet
for drug repurposing
Real World Evidence RWE
Lab: Find new use for existing drug
with AI
Module 8: AI for Biologics - Week 11*
Protein design:
Antibodies, Enzymes
AlphaFold-Multimer for
antibody-antigen
mRNA sequence optimization
Peptide drug design
Module 9: MLOps for Regulated Pharma - Week 12*
FDA/ICH guidelines for AI in drug development
Model validation + documentation
Data governance, FAIR data
Cloud: AWS HealthLake, Google
Vertex AI
Module 10: Capstone Project - Week 13-14*
*Pick 1 End-to-End Project*
1. *AI Anti-Cancer Drug*:
Target →
Generate →
ADMET →
Report
2. *Drug Repurposing for Diabetes* using knowledge graph
3. *Antibiotic Discovery* against resistant bacteria
Deliverable: Jupyter notebook + Report + Presentation
---
Tech Stack You’ll Master
Area Tools
Programming
Python,
PyTorch,
TensorFlow
Cheminformatics**
RDKit
DeepChem,
Chempro**
AI Models
GNNs,
Transformers,
Diffusion,
AlphaFold
Docking
AutoDock Vina,
DiffDock
**Databases**
ChEMBL,
PDB,
BindingDB
Best FREE Resources
1. DeepChem Tutorials
Full code for ADMET,
GNNs
2. Coursera
AI for Medicine_ - http://DeepLearning.AI
3. Stanford CS25
AI in Healthcare_ on YouTube
4.MIT
Introduction to Drug Discovery_ lectures
5.Google Colab
Free GPUs for training
Career Roles After
AI Drug Discovery Scientist, Computational Chemist,
Bioinformatics,
ML for Pharma,
R&D Data Scientist
No comments