Quantitative structure-activity relationship analysis using deep learning based on a novel molecular image input technique
Quantitative structure-activity relationship analysis using deep learning based on a novel molecular image input technique
复制标题
DOI：
10.1016/j.bmcl.2018.08.032
复制
发表时间：
2018-11-01
影响因子：
2.7
通讯作者：
Uesawa, Yoshihiro
中科院分区：
文献类型：
--
作者：
Uesawa, Yoshihiro
定量构效关系（QSAR）分析使用从分子几何计算的结构、量子化学和物理化学特征作为预测生理活性的解释变量。近年来，基于先进人工神经网络的深度学习在QSAR研究领域表现出优异的表现。虽然它具有直接从分子结构中计算特征值的特征表示学习特性，但这种势函数在QSAR建模中的使用受到限制。本研究通过将分子构象的360度图像纳入深度学习，将特征表示学习的功能应用于QSAR分析。因此，我成功构建了一个高度通用的鉴定模型，用于诱导线粒体膜电位破坏的化合物，其外部验证区域在接收器工作特征曲线>= 0.9下。
Quantitative structure-activity relationship (QSAR) analysis uses structural, quantum chemical, and physicochemical features calculated from molecular geometry as explanatory variables predicting physiological activity. Recently, deep learning based on advanced artificial neural networks has demonstrated excellent performance in the discipline of QSAR research. While it has properties of feature representation learning that directly calculate feature values from molecular structure, the use of this potential function is limited in QSAR modeling. The present study applied this function of feature representation learning to QSAR analysis by incorporating 360 degrees images of molecular conformations into deep learning. Accordingly, I successfully constructed a highly versatile identification model for chemical compounds that induce mitochondrial membrane potential disruption with the external validation area under the receiver operating characteristic curve of >= 0.9.