Representation And Manipulation Of 2D Molecular Structures
Many organisations maintain databases of chemical compounds. Some of these databases are publicly accessible; others are proprietary. They may contain a very large number of compounds; several hundred thousand is common, and some hold millions of entries. It is nevertheless possible to search for structures and data in these databases and have the results returned within seconds. A recent development involves the creation of databases containing virtual molecules. These are compounds that do not exist yet (so far as the creator is aware) but which could be synthesised readily, often using combinatorial chemistry techniques. The sizes of these virtual collections can run to billions of structures. In this chapter we introduce some of the methods used for the computational representation of molecular structures and the creation of structural databases. The procedures used to accomplish common tasks such as substructure searching are considered, including the use of graph theoretic methods, bitstring screens and the use of canonical structure representations. Finally, we give a brief overview of reaction databases and the use of Markush representations in chemical patents. The focus of this chapter will be on “2D representations” that are concerned with the chemical bonding between atoms rather than their 3D structures (the subject of Chapter 2).
KeywordsMolecular Graph Subgraph Isomorphism Structure Query Language Chemical Markup Language Chemical Abstract Service
Unable to display preview. Download preview PDF.