A structure-based framework for selective inhibitor design and optimization | JoVE Visualize

Area of Science:

Computational chemistry
Drug discovery
Artificial intelligence in medicine

Background:

Structure-based drug design (SBDD) utilizes target structures to create effective compounds.
Deep generative models aid structure-specific molecular generation but face limitations with pharmaceutical data, leading to suboptimal properties and unstable conformations.
Existing methods often neglect binding pocket interactions and struggle with selective inhibitor design.

Purpose of the Study:

To introduce Coarse-grained and Multi-dimensional Data-driven molecular generation (CMD-GEN), a framework addressing limitations in current structure-specific molecular generation.
To enhance the generation of drug-like molecules by integrating ligand-protein complex information and coarse-grained pharmacophore points.
To improve the stability and selectivity of generated molecules for drug discovery.

Main Methods:

CMD-GEN employs a hierarchical architecture to decompose 3D molecule generation into pharmacophore point sampling, chemical structure generation, and conformation alignment.
It utilizes coarse-grained pharmacophore points sampled from a diffusion model to enrich training data and bridge ligand-protein complexes with drug-like molecules.
The framework incorporates multi-dimensional data-driven approaches for molecular generation within binding pockets.

Main Results:

CMD-GEN demonstrates superior performance compared to existing methods in benchmark tests, effectively controlling drug-likeness.
The framework successfully generated molecules for three synthetic lethal targets, showcasing its applicability in complex drug discovery scenarios.
Wet-lab validation with PARP1/2 inhibitors confirmed CMD-GEN's potential for designing selective inhibitors.

Conclusions:

CMD-GEN offers a robust framework for structure-specific molecular generation, overcoming data limitations and improving molecular properties.
The hierarchical approach mitigates instability issues and enhances the generation of drug-like molecules with favorable conformations.
CMD-GEN shows significant promise for advancing selective inhibitor design in drug discovery, as evidenced by experimental validation.