IMPLEMENTATION OF ARTIFICIAL-INTELLIGENCE FOR AUTOMATIC DRUG DESIGN .1. STEPWISE COMPUTATION OF THE INTERACTIVE DRUG-DESIGN SEQUENCE

Ln this article we represent the development of an artificial-intelligence-based method for the automatic design of valid chemical structures (AISD). The key feature of the proposed algorithm is its ability to mimic many decision-making processes carried by the human drug designer during a design session. The manual drug-design process is analyzed and transformed into a computerized form by associating a weight factor with each term. These weights enable the translation of the drug designer's intution into probabilities that control the flow of the design process. The input required to initiate a design session might be as minimal as the geometry of a previously existed pharmacophoric model, up to the three-dimensional geometry of the host receptor. A design application is demonstrated by the implementation of the proposed algorithm for the design of new potent sweeteners. (C) 1994 by John Wiley and Sons, Inc.