Prediction of Hydrophobic (Lipophilic) Properties of Small Organic Molecules Using Fragmental Methods: An Analysis of ALOGP and CLOGP Methods

Abstract

Molecular hydrophobicity (lipophilicity), usually quantified as log P (the logarithm of 1-octanol/water partition coefficient), is an important molecular characteristic in drug discovery. ALOGP and CLOGP are two of the most widely used methods for the estimation of log P. This work describes an extensive reparametrization of the atomic log P values and a detailed comparison of the performance of ALOGP and CLOGP methods on the Pomona Medchem database. Only the “star list” compounds having precisely measured log P values were used in this analysis. While the overall results with both methods are similar, analysis shows that the CLOGP method is better for very small molecules in the range of 1−20 atoms. The two methods are almost comparable in the range of 21−45 atoms, while the ALOGP method has better accuracy for molecules with more than 45 atoms. Although the rms deviation and the correlation coefficient for CLOGP predictions were marginally better than those for corresponding ALOGP predictions, the latter showed a very stable performance for all classes of organic compounds analyzed. The ALOGP method can be used to compute estimates of most neutral organic compounds having C, H, O, N, S, Se, P, B, Si, and halogens. It also covers most zwitterionic compounds having amine and carboxylic acids and ammonium halide salts. The CLOGP method has improved considerably over the years to cover most neutral organic compounds, but it still has some undefined fragments. Finally, unlike CLOGP and other methods of predicting lipophilicity, the ALOGP method has multiple uses, such as the estimation of local hydrophobicity, the visualization of molecular hydrophobicity maps, and the evaluation of hydrophobic interactions in protein−ligand complexes.