Quantum mechanical molecular electrostatic potential surfaces and neural ne
tworks are combined to predict the binding energy for bioactive molecules w
ith enzyme targets. Computational neural networks are employed to identify
the quantum mechanical features of inhibitory molecules that contribute to
binding. This approach generates relationships between the quantum mechanic
al structure of inhibitory molecules and the strength of binding. Feed-forw
ard neural networks with back-propagation of error are trained to recognize
the quantum mechanical electrostatic potential at the entire van der Waals
surface of a group of training molecules and to predict the strength of in
teractions between the enzyme and novel inhibitors. Three enzyme systems ar
e used as examples in this work: AMP (adenosine mono phosphate) nucleosidas
e, adenosine deaminase, and cytidine deaminase. Quantum neural networks ide
ntify critical areas on inhibitor potential surfaces involved in binding an
d predict with quantitative accuracy the binding strength of new inhibitors
. The method is able to predict the binding free energy of the transition s
tate, when trained with less tightly bound inhibitors. The application of t
his approach to the study of enzyme inhibitors and receptor agonists would
permit evaluation of chemical libraries of potential bioactive agents. (C)
2000 John Wiley & Sons, Inc.