In an overview, the relationships between activity, property, and chemical structure of the anthracycline derivatives, as topoisomerase II enzyme inhibitors, were studied in pairs. These analogs were investigated in three categories with known half-maximal inhibitory concentration ( IC50), known octanol/water partition coefficient (logP), and both IC50 and logP unknown values. The QSAR and QSPR models were developed on analogs of the first and second groups, respectively. LogP was related to calculated pIC50 ((pIC50)calc) surprisingly (R2 = 0.929, Q2 LOO = 0.910, RMSE = 0.235, F = 157.511). External verification was performed through comparing calculated logP ( logPcalc) with three programs using 77 analogs out of the training and test sets. In application domain of the model, the suggested model had the highest correlation with ClogP method (R = 0.744) and this was better than the correlation between milogP–ClogP (R = 0.724), AlogP–ClogP (R = 0.662), and AlogP–milogP (R = 0.691). Furthermore, in the case of analogs possessing a long carbon chain substitution (nc ≥ 3), estimation methods had a significant error in calculation logP, while our suggested model provides more accurate results. The proposed algorithm can be useful in primary screening among anthracycline derivatives to select potent drug candidates.
