Natural language processing-based acquisition and analysis of information of drug metabolism

点击翻译按钮获取中文摘要

Drug metabolism is an important factor determining disposition behavior of the drug in its action sites. Moreover, genetic epigenetic difference of activity of drug metabolizing enzymes causes inter-individual variability in drug responses. Therefore, it is an important issue to predict how drugs or new chemical entities are metabolized. Although experts in this field have been manually collecting and analyzing information of drug metabolism, there have been few databases covering dozens of drug metabolizing enzymes and their interaction with chemicals. The present study was initiated to obtain the knowledge of structure-activity relationship of drug metabolism through comprehensive information acquisition from the literatures. I developed a natural language processing system for collecting information on chemical-enzyme interaction, which comprises dictionaries of chemicals and rule bases for retrieval of unregistered chemicals. The system can also detect how the chemicals interact with enzymes, simply by defining the names of the enzymes. In addition, I developed an algorithm of simultaneous analysis and visualization of structure-activity relationship towards multiple drug metabolizing enzymes. These methods would help us to proceed drug discovery and development effectively.