USE OF PARALLEL PROCESSING IN THE STUDY OF PROTEIN - LIGAND-BINDING

We have undertaken an energy minimization study of the binding of a small ligand, water, to amino acid side‐chains. These calculations have been performed on a Meiko Computing surface containing a small number of transputers and on a larger transputer array installed at Edinburgh University. The aim of this study is twofold. First, we wished to compare these potential energy maps for a given side‐chain as a function of the local secondary structure and also of the neighboring residue types. Secondly, the energy maps are found to be in good agreement with experimental distributions on the binding of water molecules to amino acid side‐chains. The use of our in‐house Meiko computing surface increases the speed of these calculations by a factor of about 25 over a VAX 11/750. With larger arrays of transputers, such as that at Edinburgh “concurrent” Supercomputer Project, we achieve increases in speed of over 200 such that these energy maps can be calculated at interactive speeds. Copyright © 1990 John Wiley & Sons, Inc.