Sampling a trillion-sized GDB-20 database of drug-like molecules by generative artificial intelligence.
- 1Department of Chemistry, Biochemistry and Pharmaceutical Sciences, University of Bern Bern 3012 Switzerland [email protected].
Abstract
Identifying new chemical entities for small molecule drug discovery requires the continued exploration of its vast chemical space, which has been speculated to contain 10⁶⁰ or more molecules. Here we used generative artificial intelligence to sample a GDB-20 database estimated to contain 32 trillion organic molecules corresponding to 26.9 million molecular graphs up to 20 nodes selected for their limited topological complexity. We show that GDB-20 populates a distinct chemical space of sp³-rich, 3D-shaped and natural product-like molecules with favorable synthetic accessibility score. The relevance for drug discovery of our 12 billion sample, named GDB-20s and corresponding in size to approximately 0.04% of the estimated full database, is suggested by high compliance with drug-likeness and toxicity filters, and the coverage of thousands of scaffolds occurring in bioactive molecules recorded in the ChEMBL database. The generative models used to produce GDB-20s, together with the 12-billion-molecule collection are available for download at https://gdb.unibe.ch.
Related
- GEOMIND: a hybrid generative artificial intelligence model for geopolymer design and optimization
- From Generative Molecular Design to Agentic Artificial Intelligence in Drug Discovery
- Generative and Foundation-Model Artificial Intelligence in Food Safety Analysis, Authenticity Identification, and Quality Control: A Systematic Review of Validation and Decision Readiness.
- VIBE-QC: an open-source, general-purpose electronic-structure code for molecules and solids developed with generative AI
- Recovering molecules from coarse-grained beads: free-energy-conditioned generative backmapping across chemical space
BibTeX
@article{buehler2026sampling,
title = {{Sampling a trillion-sized GDB-20 database of drug-like molecules by generative artificial intelligence.}},
author = {Ye Buehler and Sacha Javor and Jean-Louis Reymond},
journal = {Digital discovery},
year = {2026},
month = oct,
doi = {10.1039/d6dd00213g},
url = {https://europepmc.org/article/MED/42948043},
note = {CC BY},
}
RIS
TY - JOUR
TI - Sampling a trillion-sized GDB-20 database of drug-like molecules by generative artificial intelligence.
AU - Ye Buehler
AU - Sacha Javor
AU - Jean-Louis Reymond
PY - 2026
DA - 2026/10/16
JO - Digital discovery
DO - 10.1039/d6dd00213g
UR - https://europepmc.org/article/MED/42948043
AB - Identifying new chemical entities for small molecule drug discovery requires the continued exploration of its vast chemical space, which has been speculated to contain 10⁶⁰ or more molecules. Here we used generative artificial intelligence to sample a GDB-20 database estimated to contain 32 trillion organic molecules corresponding to 26.9 million molecular graphs up to 20 nodes selected for their limited topological complexity. We show that GDB-20 populates a distinct chemical space of sp³-rich, 3D-shaped and natural product-like molecules with favorable synthetic accessibility score. The relevance for drug discovery of our 12 billion sample, named GDB-20s and corresponding in size to approximately 0.04% of the estimated full database, is suggested by high compliance with drug-likeness and toxicity filters, and the coverage of thousands of scaffolds occurring in bioactive molecules recorded in the ChEMBL database. The generative models used to produce GDB-20s, together with the 12-billion-molecule collection are available for download at https://gdb.unibe.ch.
SN - 2635-098X
ER -