An application of QSRR approach and multiple linear regression method for lipophilicity assessment of flavonoids.

The analysis of quantitative structure-retention relationships (QSRR) is useful tool for assessment of compound's lipophilicity/hydrophobicity due to similarity between its retention in chromatographic system and ability to permeation through biological membranes. The main goal of this study was to compare usefulness of two reversed-phase chromatographic columns (Synergy POLAR and Synergy-FUSION) for lipophilicity assessment of 30 structurally diverse flavonoids using the QSRR approach and multiple linear regression method. The developed MLR models included the mechanistically interpretable geometrical descriptors: 3D Molecule Representation of Structure based on Electron diffraction (3D-MoRSE) and Radial Distribution Function (RDF). Both models were evaluated by the internal and external validation and selected descriptors were further interpreted. According to obtained results the FUSION-RP column can be recommended to log k prediction of flavonoids. The comprehensive interpretation of molecular descriptors was used to present the molecular mechanisms and structural features governing the chromatographic retention of tested compounds.