期刊
JOURNAL OF CHEMICAL INFORMATION AND MODELING
卷 62, 期 23, 页码 5907-5917出版社
AMER CHEMICAL SOC
DOI: 10.1021/acs.jcim.2c009822022
关键词
-
类别
资金
- National Key R&D Program of China [2020YFB0204803]
- Guangdong Introduc-ing Innovative and Entrepreneurial Teams [2016ZT06D211]
- Guangzhou ST Research Plan [202007030010]
A novel framework called DRlinker was proposed in this study to control fragment linking and generate compounds with specified attributes through reinforcement learning. The method has been demonstrated to be effective in various tasks, including controlling linker length and log P, optimizing predicted bioactivity of compounds, and achieving various multiobjective tasks.
Fragment-based drug discovery is a widely used strategy for drug design in both academic and pharmaceutical industries. Although fragments can be linked to generate candidate compounds by the latest deep generative models, generating linkers with specified attributes remains underdeveloped. In this study, we presented a novel framework, DRlinker, to control fragment linking toward compounds with given attributes through reinforcement learning. The method has been shown to be effective for many tasks from controlling the linker length and log P, optimizing predicted bioactivity of compounds, to various multiobjective tasks. Specifically, our model successfully generated 91.0% and 93.9% of compounds complying with the desired linker length and log P and improved the 7.5 pChEMBL value in bioactivity optimization. Finally, a quasi-scaffold-hopping study revealed that DRlinker could generate nearly 30% molecules with high 3D similarity but low 2D similarity to the lead inhibitor, demonstrating the benefits and applicability of DRlinker in actual fragment-based drug design.
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