期刊
ANGEWANDTE CHEMIE-INTERNATIONAL EDITION
卷 60, 期 35, 页码 19477-19482出版社
WILEY-V C H VERLAG GMBH
DOI: 10.1002/anie.202104405
关键词
de novo design; deep learning; drug discovery; neural network; nuclear receptor
资金
- Swiss National Science Foundation [205321_182176]
- RETHINK initiative at ETH Zurich
- Novartis Forschungsstiftung (FreeNovation grant AI in Drug Discovery)
- Projekt DEAL
Chemical language models combined with the beam search algorithm as an automated molecule design and scoring technique can generate novel compounds with potential bioactivity. The newly discovered inverse agonists can be synthesized in a few simple reaction steps and exhibit low micromolar to nanomolar potency towards RORg. This model-intrinsic sampling technique eliminates the strict need for external compound scoring functions, further expanding the applicability of generative artificial intelligence to data-driven drug discovery.
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 RORg. 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.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据