Should I generate ionization and tautomeric states when adding molecules to my Phase database?
There is no recommended procedure for generating ionization and tautomeric states for a Phase database, because the decision depends on the specific molecules and their intended purpose. Any decision should take into account the following issues: • The Phase feature definitions can account for most common ionic forms even if the structures are neutral: the SMARTS patterns for ionizable groups can be included in the relevant positive ion or negative ion feature definitions. • Generating ionization and tautomeric states can provide better coverage of the possible states, but can significantly increase the disk space used. • Inclusion of tautomeric states can result in unusual hits, such as a donor mapped to an enol. These possibilities might require the addition of more patterns to the feature definitions to exclude unreasonable hits. The CAC database, distributed with Phase, includes ions and tautomers generated at pH 7 ± 2, with the two most stable states retained. Including 4 tautomer