[摘要]：The anal. of chem. diversity has become a topic of considerable interest in recent years. This interest has been stimulated by the challenge of discovering novel small-mol. pharmaceuticals. The development of technologies such as combinatorial synthesis and high-throughput screening has made it possible to explore drug-like regions of chem. space in relatively short times. Chem. space is vast and the problem of selecting which region of that space to explore remains a key issue in drug discovery. By analyzing the scaffold content of the CAS Registry, the authors attempt to characterize in a comprehensive way the structural diversity of org. chem. The scaffold of a mol. is taken to be its framework, defined as all its ring systems and all the linkers that connect them. Framework data from more than 24 million org. compds. are analyzed. The distribution of frameworks among compds. is found to be top-heavy, i.e., a small percentage of frameworks occurs in a large percentage of compds. When frameworks are analyzed at the graph level, an even more top-heavy distribution is found: half of the compds. can be described by only 143 framework shapes. The most significant finding is that the framework distribution conforms almost exactly to a power law. This suggests that the more often a framework has been used as the basis for a compd., the more likely it is to be used in another compd. This may be explained by the cost of synthesis: making a new deriv. of a framework is probably less costly if many other derivs. are known. The authors believe that this power law is evidence that the minimization of synthetic cost has been a key factor in shaping the known universe of org. chem.
