Development of an AI(Artificial Intelligence)-Based Optimization System for Tandem Mass Spectrometry

摘要

The interfacing of the expert system to the triple quadrupole mass spectrometer (TQMS) proved the value of expertise, encoded in the form of rules, to complex optimization problems. The system is able to optimize the output of a complex instrument running chemical compounds in a way not practical with manual methods. Because the expert system approach allows the instrument to be tuned quickly, multiple mass range, or even individual mass pair tuning is now practical, resulting in large gains in instrument sensitivity. Two significant problems with the application of knowledge based systems to chemical instrumentation have been encountered. First, there is a significant learning curve associated with applying the technology, and second, the knowledge based systems software tools and the supporting hardware are expensive. While our experience has shown that the hardware and software tools are cost effective for developing the system, these high costs make fielding multiple copies of the expert system, using the total development environment, economically prohibitive. The alternative is to port the development system to other hardware using conventional languages, but this approach is practical only where the large costs of porting the software can be amortized over many systems. 8 refs., 16 figs. (ERA citation 12:009518)