Structure-based drug design
Designing drug molecules by exploiting the 3D structure of a biological target — using shape and chemistry of the binding site to guide which compounds to synthesize or screen.
Structure-based drug design (SBDD) uses experimentally determined or predicted 3D structures of a protein (or other biological target) to guide the design of compounds that bind to it. If you know the shape and chemical properties of a protein’s binding pocket, you can design or select molecules whose shape and charge distribution are complementary — like a key shaped to fit a specific lock.
The workflow typically involves:
- Obtaining a target structure (X-ray crystallography, cryo-EM, NMR, or — increasingly — AlphaFold prediction)
- Identifying the binding site (experimentally known, or computationally predicted)
- Docking candidate molecules to score predicted binding
- Iterating: synthesizing or purchasing top candidates, testing them, using the resulting structure-activity data to refine the design
How AI has accelerated SBDD:
AlphaFold 2 and 3 have made structure-based approaches possible for proteins where no experimental structure exists, dramatically expanding the target space. AI scoring functions can evaluate binding poses faster and sometimes more accurately than classical physics-based scoring. Isomorphic Labs, Schrödinger, and others have built proprietary platforms integrating generative design directly into SBDD workflows.
Key limitation: crystal structures capture one conformation of a protein. Proteins are flexible, and the conformation in a crystal may not represent the dominant conformation in solution or in the cellular environment where a drug would actually act. Accounting for protein flexibility (induced fit docking, molecular dynamics) adds complexity and compute cost.
Related terms: Virtual Screening, De Novo Design
Related guide: Health & Medicine
Related tool: AlphaFold