论文部分内容阅读
在化学反应处理的计算模型的基础上 ,完成了一个实验性的目标化合物析分系统 (TargetParsingSystem ,TPS)。目前它主要用于测试和优化反应规则的选择和表述方法。TPS的工作逻辑基本上模仿人的思维过程 ,它通过对目标分子析分找出谋略键 ,根据谋略键将目标分子分拆成前体 ,再通过对原料库的检索确定是否已是可得原料等 3个步骤完成反合成分析。通过对一些小分子和天然产物的测试 ,表明它已能给出令人满意的反合成分析结果。由于TPS在知识量上的局限性 ,对复杂化合物的析分还不可能达到实用水平。但是 ,可以预见通过对知识库的扩充和对TPS的不断完善 ,它有望发展成为一个真正意义上的计算机辅助合成设计系统。
Based on the computational model of chemical reaction processing, an experimental Target Compounding System (TPS) was completed. It is currently mainly used to test and optimize the selection and presentation of reaction rules. The working logic of TPS basically imitates the human thinking process. It divides the target molecule into precursors according to the strategy key by analyzing the target molecules, and then determines whether the raw materials are available by searching the raw material bases 3 steps to complete the anti-synthesis analysis. Tests of some small molecules and natural products show that it has been able to give satisfactory results in reverse synthesis. Due to the limited knowledge of TPS, the resolution of complex compounds is unlikely to reach a practical level. However, it is foreseeable that through the expansion of knowledge base and continuous improvement of TPS, it is expected to develop into a truly computer-aided synthesis design system.