W. Jeffrey Howe

3D database searching and de novo construction methods in molecular design

Abstract Computer-based lead finding algorithms which attempt to design, on a de novo basis, ligands that will complement a known receptor site cavity face some major problems in terms of a combinatorial design space and the synthesizability of the designed molecules. On the other hand, typical 3D database search methods provide a different set of challenges. Both of these approaches are ultimately pointed toward the same goal and can be used together productively. In this article we describe advances in both areas: we first describe extensions to our de novo ligand design software which combines (a) a tree-based conformational search over a library of fragments, and (b) a form of simulated annealing which allows designed ligands to crawl around the binding site even as their structures are changing. In the second part, we then discuss an implementation of the database approach which allows users to formulate 3D substructure, superstructure, or similarity queries based upon demonstrated or hypothetical requirements for activity. Finally, we draw the two approaches together with an example of current research interest, showing how one method can feed the other.

A fast algorithm for generating smooth molecular dot surface representations

The smooth molecular surface originally described by Richards and later implemented by Connolly in his MS program has become an important visualization technique in the field of molecular modeling. We describe here a new algorithm, called USURF, which approximates the MS dot surface, but with a twofold to sixfold enhancement of speed. The algorithm has been incorporated into our interactive modeling system, Mosaic, and is also available as a stand-alone program.

Computer Modeling of Constrained Peptide Systems

Three approaches are described for dealing with conformational problems in constrained peptide systems. The first approach addresses the problem of de novo design of peptides where the structure of the binding or active site of the target protein is known. The method is based on a procedure, called GROW, which links together amino acid fragments from a library of amino acid conformations such that an “optimum fit” and solvation energy for the ligand are obtained. The second approach addresses the problem of modeling the surface loops of proteins. It is based upon the novel use of simulated annealing within the framework of a “random-tweak” type procedure [Shenkin et al., Proteins, 1989]. The method has been tested on 6- and 8-residue loops, and yields results that agree well with experiment at relatively little computational cost. The third approach addresses both cyclic peptides and surface-loop modeling using a mass-weighted molecular dynamics (MWMD) procedure. It is shown that MWMD is clearly superior to normal-mass MD sampling at equivalent temperatures, and is numerically stable. Moreover, the sampling of individual torsional degrees of freedom is also seen to be essentially complete for the case studied here, viz. for pressinoic acid, a cyclic hexapeptide derived from vasopressin.

Intermediary-Free Substructure Searching

For the past few years, Upjohns Cousin system has been in use for chemical registry, substructure searching (SS), biological data retrieval, and report generation. This discussion will focus on the capabilities of the SS portion of the system and on our experiences in designing and releasing a graphics-based system intended to be operated directly by scientist end-users. Some of the trade-offs involved in the “make” versus “buy” decision are also discussed.