Tautomer enumeration

Successful virtual screening relies on an appropriate and correct description of the molecules being screened two particular sources of error are protonation state and tautomeric form. Oellien et have developed an exhaustive tautomer enumeration approach based on a set of 21 predefined transforms. The group do not attempt to say which tautomer is favoured, only that every form is represented. Databases built using all tautomers show better performance especially when screened with pharmacophore queries, as the bias introduced by most compound registration tools as to the correct tautomer is removed. 

MiDetti, F., Storchi, L., Sforna, G., Cross, S., and Crudani, G. (2009) Tautomer enumeration and stability prediction for virtual screening on large chemical databases. Journal of Chemical Information and Modeling, 49, 68-75. 

An apparently trivial but important point to be considered is the generation of structures. Although for simple systems all the possible tautomeric forms can be easily drawn, in other cases the situation is less clear for example, 15 sttuctures for xanthine were provided to the participants in the SAMPL2 challenge as many as 40 sttuctures (tautomers and conformers) have been considered for warfarin [8]. Several tools exist to handle the tautomer enumeration problem [9], but these rely on experimental data and/or empirical rules (mosdy pJC values) and hence not all types of tautomerism might be described correctly. 

One element of database generation that is a key consideration is whether to expand the representative compounds to include alternative tautomers, protonated and deprotonated forms of the molecule, and also to enumerate stereochemistry fully if not specified in the input. Depending on the molecules in question and the options considered, these can lead to a 10-fold increase in the size of the database to be explored. However, such an expansion is necessary if methods are used that are sensitive to such chemical precision (e.g., docking). For 3D similarity searching, it is sometimes more efficient to consider various modifications to the query, leading to multiple searches against a smaller database. 

The transformation used above to enumerate tautomers would lead to identical products when applied to symmetrically substituted pyrazoles. The set of structures generated in the enumeration process is converted to a sorted list of canonical SMILES [23] from which duplicates are easily eliminated. Structures registered in alternative tautomeric forms are converted to identical lists of SMILES that can each be represented by their common first member. This effectively extends the definition of canonical SMILES to cover an ensemble of tautomeric forms and makes it possible to check for duplicate structures without having to register multiple forms [16, 26]. 

Sayle, R. and Delany, J. Canonicalization and enumeration of tautomers. EuroMug-99. Cambridge, UK, 28-29 October 1999, http //www.daylight. 

Sayle, R. 1999. Canonicalization and Enumeration of Tautomers, http //www.dayhght.com/meetings/emug99/ Delany/taut html/index.htm. 

Definition of unambiguous business rules is not always straightforward and in these situations it is useful to generate an ensemble of equivalent structures.If rules to rank the structures in the ensemble can be defined, then it becomes possible to select one of these to represent the ensemble. For example, enumeration of tautomers using Leatherface creates a sorted list of canonical SMILES, the first of which is used to represent the ensemble. This effectively extends the concept of canonicalisation to an ensemble of structures, and the term canonical tautomer may be encountered in the literature. ... 

---