Renal clearance is one of the main pathways for a drug to be cleared from plasma. The aim of this study is to develop in-silico models to find out the relationship between the type of renal clearance, and structural parameters.Literature data were used to categorise the drugs into those that undergo tubular secretion and those that undergo reabsorption. Different structural descriptors (VolSurf descriptors, Abraham solvation parameters, data warrior descriptors, logarithm of distribution coefficient at pH = 7.4 (logD7.4)) were applied to develop a mechanistic model for estimating renal clearance class whether its secretion or reabsorption.The results of this study show that logD7.4 and the number of hydrogen bond donors, as well as available uncharged species (AUS7.4), are the most effective descriptors to establish mechanistic models for predicting renal clearance class. The classification models were established with a level of accuracy of more than 75%.Developed models of this study can be helpful to predict renal clearance class for new drug candidates with an acceptable error. Hydrophilicity and hydrogen bond formation ability of drugs are among the main descriptors.
Keywords: Logistic regression; model; prediction; renal clearance; structural descriptors.