Ionization and Tautomer Model

Preparation of virtual screening databases starts with standardization of the input SMILES. This procedure was originally developed to deal with databases from commercial suppliers. Preferred tautomeric forms are generated in this step and ionized species are neutralized. Ionization states are set in the second step for biased equilibria and multiple forms are enumerated in a third step to represent balanced equilibria. The model treats an equilibrium as balanced if the equilibrium constant associated with its defining rule is likely to be less than about 1.5 log units. 

Our current ionization and tautomer model encodes biased and balanced equilibria with 108 and 35 rules respectively. Extensive use is made of atom typing to specify chemistry in as generic a manner as possible, which reduces the number of rules required, for example, a number of heteroaromatic nitrogen types, including selected 2-aminopyridines and imidazoles, are aggregated into a definition of bases with pK  

---