Graph applications in chemoinformatics and structural bioinformatics

INTRODUCTION Chemistry space is exceedingly large. Recent estimates put the number of potentially 'drug-like' molecules at anything between 10 12 and 10 180 (Gorse, 2006). The overwhelming majority of these molecules never has been, and never will be, synthesized but methods are nevertheless required to determine which of these potential compounds should be made. Some large pharmaceutical/ agrochemical companies maintain corporate databases of millions of molecules. The discovery of New Chemical Entities (NCEs) which may become drugs depends on the successful mining ABSTRACT The focus of this chapter will be the uses of graph theory in chemoinformatics and in structural bio-informatics. There is a long history of chemical graph theory dating back to the 1860's and Kekule's structural theory. It is natural to regard the atoms of a molecule as nodes and the bonds as edges (2D representations) of a labeled graph (a molecular graph). This chapter will concentrate on the algorithms developed to exploit the computer representation of such graphs and their extensions in both two and three dimensions (where an edge represents the distance in 3D space between a pair of atoms), together with the algorithms developed to exploit them. The algorithms will generally be summarized rather than detailed. The methods were later extended to larger macromolecules (such as proteins); these will be covered in less detail. of the information stored in these databases. Such information may be explicit (e.g. the molecules may be annotated with chemical reaction or activity data) or may be implicit in the structure of a molecule. For many years isomorphism algorithms have formed the basis of structural comparison between pairs of molecules in these databases, designed to extract this implicit information. The main purpose of this chapter is to introduce the concept of chemoinformatics to practitioners from the field of graph theory and to demonstrate the widespread application of graph-theoretic techniques to the solving of chemoinformatics problems. However, many graph-theoretic algorithms from chemoinformatics have subsequently been adapted for the structural comparison of macromolecules in the field known as structural bioinformatics. A secondary aim of the chapter is to provide a brief overview of these applications. The layout of the chapter will now be described. In the Background section some definitions of chemoinformatics are given and the topic is placed within the context of the drug discovery process. Structural bioinformatics is also defined and the necessary graph theoretic notation is introduced. The two main …