作者: Robert D. Brown , Yvonne C. Martin
DOI: 10.1021/CI9501047
关键词: Selection method 、 Selection (genetic algorithm) 、 Pharmacophore 、 Artificial intelligence 、 3d descriptors 、 Hierarchical clustering 、 Pattern recognition 、 Mathematics 、 Structure based 、 Cluster analysis 、 Structure (mathematical logic) 、 Data mining
摘要: An evaluation of a variety structure-based clustering methods for use in compound selection is presented. The MACCS, Unity and Daylight 2D descriptors; 3D rigid flexible descriptors two in-house based on potential pharmacophore points, are considered. Ward's group-average hierarchical agglomerative, Guenoche divisive, Jarvis−Patrick nonhierarchical compared. results suggest that best at separating biologically active molecules from inactives, prerequisite good method. In particular, the combination MACCS was optimal.