Work

Sampling a trillion-sized GDB-20 database of drug-like molecules by generative artificial intelligence

Year: 2026

Type: article

Abstract: Generative AI sampled 12 billion GDB-20 molecules, revealing an sp 3 -rich, 3D-shaped and natural product-like chemical space with potential synthetic accessibility, drug-likeness, toxicity-filter compliance and broad bioactive scaffold coverage.

Source: Digital Discovery

DOI: 10.1039/d6dd00213g

Authors: Ye Buehler, Sacha Javor, Jean-Louis Reymond

Institution: University of Bern

Language: English

Topic: Computational Drug Discovery Methods

Subfield: Drug Discovery

Field: Pharmacology, Toxicology and Pharmaceutics

Domain: Life Sciences

Open Access status: diamond

Funders: National Science Foundation, University of Bern, Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung

Award: Chemical Space Design of Small Molecules and Peptides