Our approach and deliverables
How does chemical-space based drug discovery work?

- Reported 3D structures of targets alone or existing drug-target interactions
- Experimental data (e.g., co-crystallization)


- Choice of models whether data is abundant (e.g generative models) or scarce (e.g Genetic algorithm)
- Similarity models (e.g., bioactivity prediction, on-/off- target prediction)
- Predictive models (e.g. for properties)
- Physics-based models (e.g., docking)


- Molecular descriptors (e.g., shape, pharmacophores)
- Experimental data
(e.g., bioactivity, toxicity)
Our Workflow
