2006 American Chemical Society Award for Computers in Chemical and Pharmaceutical Research - Of molecules and humans
2006 American Chemical Society Award for Computers in Chemical and Pharmaceutical Research - Of molecules and humans
复制标题
DOI:
10.1021/jm0608964
复制
发表时间:
2006-11-02
影响因子:
7.3
通讯作者:
Gasteiger, Johann
中科院分区:
文献类型:
--
作者:
Gasteiger, Johann
The understanding and modeling of the relationships between the structure of a molecule and its biological activity are the central themes of medicinal chemistry. Clearly, the biological activity of a compound cannot directly be calculated by first principles from the molecular structure. In this situation, an indirect approach has to be taken to find a relationship between the molecular structure of a compound and its biological activity (Figure 1). First, the structure of a molecule has to be represented by structure descriptors. Then these structure descriptors are used for modeling the biological activity. This second step asks for a set of molecules and their associated biological activities in order to apply inductive learning methods such as statistical or pattern recognition methods or artificial neural networks to establish a relationship between the structure descriptors and the biological activity.In this article we will concentrate on the first step, the representation of molecules. In the past few decades a wide variety of methods have been developed to derive structure descriptors for a molecule. 1-3 We propose here an ordering scheme that allows one to determine the level of sophistication in structure representation. In this endeavor we will compare molecules with humans, both being three-dimensional objects having surfaces, having different types of surfaces, having chiral parts, and being flexible (Figure 2). This comparison will allow us to decide what results we can expect from a certain type of molecular structure representation and thus give guidelines on which structure representation to choose for a certain problem at hand.