Competence and resources for prediction of acid-base properties (pKa)
and lipophilicities (logPo/w)
in cases when an experimental determination is inconclusive or impossible. Predictions may be based on related experimental data, purely computational values, or a lucrative combination of the two. Such possibilities can conserve time and laboratory resources infrequent cases when approximate knowledge of the value is sufficient for planning the experiment, or estimated value is good enough for the purpose. The used approaches include a multitude of computational methods
implemented in Turbomole
software, COSMO-RS method
for calculations of liquid-state properties, access to a powerful computer cluster
, as well as advanced statistical methods (e.g. PCA, Random Forest
) for creating and validating predictive models.