Schneider Gisbert
Johann Wolfgang Goethe-Universität, Beilstein Professor of Cheminformatics, Institut für Organische Chemie und Chemische Biologie, Marie-Curie Str., Frankfurt, 11, D-60439, Germany.
Curr Med Chem. 2002 Dec;9(23):2095-101. doi: 10.2174/0929867023368755.
Recent developments in combinatorial molecular design using virtual screening methods are summarised. These include similarity-based compound clustering techniques, structure-based docking and scoring, and fragment-based de novo design. Three major trends have been identified: i) the design of small target-focused compound libraries yielding activity-enriched sets of molecules; ii) advanced prediction methods for "drug-like" molecular properties complement activity predictions in the library design process, forming a multi-dimensional objective function; iii) "cherry picking" of selected products is increasingly used in lead generation and optimisation compared to purely educt-driven library design methods aiming at maximising structural diversity.
总结了使用虚拟筛选方法进行组合分子设计的最新进展。这些进展包括基于相似性的化合物聚类技术、基于结构的对接和评分，以及基于片段的从头设计。已确定了三个主要趋势：i）设计聚焦于小靶点的化合物库，产生富含活性的分子集；ii）针对“类药物”分子性质的先进预测方法在库设计过程中补充活性预测，形成多维目标函数；iii）与旨在最大化结构多样性的纯反应物驱动的库设计方法相比，在先导化合物生成和优化中越来越多地使用对选定产物的“挑选精华”。
