Classification of Current Scoring Functions
Classification of Current Scoring Functions
复制标题
DOI：
10.1021/ci500731a
复制
发表时间：
2015-03-01
影响因子：
5.6
通讯作者：
Wang, Renxiao
中科院分区：
文献类型：
--
作者：
Liu, Jie;Wang, Renxiao
评分函数是一类广泛应用于基于结构的药物设计中的计算方法，用于评估蛋白质-配体相互作用。自20世纪90年代初以来，已经发布了数十种评分函数。在文献中，评分函数通常被分类为基于力场、经验和基于知识。这一分类方案已经被引用了十多年，最近的一些出版物仍然反复引用。不幸的是，它没有反映这一领域的最新进展。此外，用于描述不同类型的评分函数的命名约定在文献中有些混乱，这可能会使该领域的新手感到困惑。在这里，我们表达了我们对当前评分函数的最新分类方案和适当命名约定的观点。我们建议，他们可以分为基于物理的方法，经验的评分功能，基于知识的潜力，和基于神经网络的评分功能。我们还概述了不同类别的评分功能之间的主要区别和联系。
Scoring functions are a class of computational methods widely applied in structure-based drug design for evaluating protein-ligand interactions. Dozens of scoring functions have been published since the early 1990s. In literature, scoring functions are typically classified as force-field-based, empirical, and knowledge-based. This classification scheme has been quoted for more than a decade and is still repeatedly quoted by some recent publications. Unfortunately, it does not reflect the recent progress in this field. Besides, the naming convention used for describing different types of scoring functions has been somewhat jumbled in literature, which could be confusing for newcomers to this field. Here, we express our viewpoint on an up-to-date classification scheme and appropriate naming convention for current scoring functions. We propose that they can be classified into physics-based methods, empirical scoring functions, knowledge-based potentials, and descriptor-based scoring functions. We also outline the major difference and connections between different categories of scoring functions.