An accurate charge model in molecular mechanics force fields ensures better logD correlations

For a drug to have an effect, it invariably has to cross the cell membrane to reach its target protein. The distribution coefficient (logD), which describes the distribution of a small molecule between different solvents, is a crucial indicator of how efficient the drug is at crossing a membrane. Though many chemoinformatics tools are available to make logD predictions, the accuracy and generalisability of these tools remains limited, especially those based on regression or machine-learning techniques.