The importance of calculating not only the correct tautomer, but also the correct protonation state and conformation in 3D modeling applications is emphasized. Above all, identifying and characterizing the most stable form of a ligand under physiological conditions is seen to be the key to successful 3D modeling. Modeling strategies that make use of the performance of modern hardware can employ physically more appropriate models than most currently in use and still be easily applicable to large numbers of compounds. Because the performance of quantitative structure–property relationships is likely to be limited by the available training and validation data, we must either find new sources of such data or resort to explicit modeling, which can partly be parameterized using definitive ab initio calculations for reference data such as gas-phase proton affinities.
|Number of pages||7|
|Journal||Journal of Computer-Aided Molecular Design|
|Publication status||Published - 2010|