COMPUTER-AIDED DRUG DESIGN

Two distinct approaches are possible in the area of computer-aided drug design. If the molecular structure of the target macromolecule is known the methods are obvious and direct and have achieved a high level of sophistication. That area may be extended by using computational techniques to predict protein structure as illustrated here by the interleukin-4 receptor. When the only lead is a set of known active compounds or knowledge of a biochemical transformation which is to be interrupted, then the path is less direct. Currently favoured tactics include the use of molecular similarity methods and the employment of neural networks. Recent advances include the prediction of the relative potency of different chiral forms of drugs.