The Core and Most Useful Molecules in Organic Chemistry

摘要

On the most abstract level, the millions of known chemicals and reactions constituting organic chemistry can be represented as a complex network,[1] in which compounds correspond to nodes and reactions to directed connections between these nodes. We have recently shown[2] that such a network has a scale-free topology similar to that of the World Wide Web and that by analyzing its time evolution, it is possible to derive statistical laws that describe and also predict how and which types of molecules are/will be synthesized. The major limitation of this statistical approach is that it provides information only about the average properties of molecules and the overall network structure - at the same time, it does not tell us how specific chemicals are arranged or what roles they play in the network of chemistry. Herein, we address these important issues with the aim of identifying molecules central to and most useful in organic synthesis. By using mathematical tools from network theory and statistical physics, we demonstrate that 1) there exists a small set of strongly connected, chemically diverse substances (the core) from which the majority of other known organic compounds (the periphery) can be made in three or fewer synthetic steps, and that 2) this central structure is surrounded by small islands that do not connect either to the core or to the periphery. Furthermore, by defining the usefulness of a compound as proportional to the number of other compounds that can be made from it, we develop Monte Carlo (MC) search algorithms to identify small optimal sets of maximally useful chemicals. Aside from purely scientific interest, the knowledge of such most useful collections of compounds should be of practical value to specialty chemical companies, which could use it to optimize product selection and cater for the most diverse group of chemical customers.