TY  - JOUR
A1  - Moret, Michael
A1  - Helmstädter, Moritz
A1  - Grisoni, Francesca
A1  - Schneider, Gisbert
A1  - Merk, Daniel
T1  - Beam search for automated design and scoring of novel ROR ligands with machine intelligence
T2  - Angewandte Chemie
N2  - Chemical language models enable de novo drug design without the requirement for explicit molecular construction rules. While such models have been applied to generate novel compounds with desired bioactivity, the actual prioritization and selection of the most promising computational designs remains challenging. Herein, we leveraged the probabilities learnt by chemical language models with the beam search algorithm as a model-intrinsic technique for automated molecule design and scoring. Prospective application of this method yielded novel inverse agonists of retinoic acid receptor-related orphan receptors (RORs). Each design was synthesizable in three reaction steps and presented low-micromolar to nanomolar potency towards RORγ. This model-intrinsic sampling technique eliminates the strict need for external compound scoring functions, thereby further extending the applicability of generative artificial intelligence to data-driven drug discovery.
KW  - de novo design
KW  - deep learning
KW  - drug discovery
KW  - neural network
KW  - nuclear receptor
Y1  - 2021
UR  - http://publikationen.ub.uni-frankfurt.de/frontdoor/index/index/docId/63925
UR  - https://nbn-resolving.org/urn:nbn:de:hebis:30:3-639259
SN  - 1521-3773
N1  - A previous version of this manuscript has been deposited on a preprint server (https://doi.org/10.26434/chemrxiv.14153408.v1).
N1  - This research was supported by the Swiss National Science Foundation (grant no. 205321_182176 to G.S.), the RETHINK initiative at ETH Zurich, and the Novartis Forschungsstiftung (FreeNovation grant “AI in Drug Discovery” to G.S.). Open access funding enabled and organized by Projekt DEAL.
VL  - 60
IS  - 35
SP  - 19477
EP  - 19482
PB  - Wiley-VCH
CY  - Weinheim
ER  -