ChemInform Abstract: SMILES Extensions for Pattern Matching and Molecular Transformations: Applications in Chemoinformatics.

The selection and modification of atoms or functional groups underly many of the manipulations central to molecular modeling. It has become even more important to automate these tasks with the current prevalence of work with large databases of molecules. We have devised SUPER-SMILES, a conceptually simple set of extensions to the SMILES line notation, whose key features are addition and deletion facilities, macros, atom tagging, disjunctions, and constraints. This superset of SMILES enables us to carry out transformations on individual molecular structures or across members of a database with a pattern-matching protocol. The principal advantage of SUPER-SMILES is the ability to specify chemical reactions with a very simple augmentation of the SMILES line notation. For example, in conjunction with macros, it is possible to represent the displacement of tosylate with phenoxy by the expression "(Delete Tosyl) (Add Phenoxy)". SUPER-SMILES thus represents a unified approach to molecular structure specification and modification and can easily be applied to large datasets of molecules. This functionality has been implemented within the PROMETHEUS suite of CAMD programs. We demonstrate its use in carrying out such operations as atom-type assignment, protonation of molecules, valency checking, and hydrogen addition. Further applications such as library design and construction immediately suggest themselves.