英文原题:Structure-based Identification of Organoruthenium Compounds as Nanomolar Antagonists of Cannabinoid Receptors
Posted by Luyao Ma on January 26, 2024
Posted in news
英文原题:Structure-based Identification of Organoruthenium Compounds as Nanomolar Antagonists of Cannabinoid Receptors
Posted by Luyao Ma on January 22, 2024
Posted in news
Posted by Luyao Ma on February 9, 2023
Posted in news
Posted by Luyao Ma on February 9, 2023
Posted in news
【VIP来稿】北生所黄牛课题组Int J Mol Sci论文:雪中送炭还是锦上添花——基于靶标结构的虚筛为新药研发贡献了什么?
Posted by Luyao Ma on November 19, 2022
Posted in news
近年来,机器学习(包括深度学习)打分函数被广泛用于预测蛋白-小分子结合自由能,并且在某些benchmark数据集上表现出优异的预测能力。但如果测试集含有与训练集相似的蛋白-小分子复合体,那么机器学习可能仅通过“学习”训练集的特定分布,就能在类似分布的测试集上取得优异表现。但这只是一种表面现象,因为这类模型往往在异分布的数据集上表现不理想,即不具备泛化能力(Generalization ability)。