The Caco-2 cell line is a standard in vitro model for assessing a compound’s absorption and permeation properties. By combining Cyprotex’s unique LIMS (LabsysTM) with state of the art automation, Cyprotex have developed a highly automated and robust procedure for the generation of high throughput Caco-2 permeability data. In a joint publication with the University of Torino, Cyprotex present how the data generated from the high throughput Caco-2 permeability assay can be used to build new predictive models based on a new tool known as block relevance analysis.
Block Relevance (BR) analysis interprets PLS (partial least squares) models based on VolSurf+ (VS+) descriptors and provides easy-to-read mechanistic insights. The VS+ descriptors are assigned by BR analysis to six blocks based on the physicochemical properties illustrated below.
Experimental Caco-2 log Papp values of the 54 training compounds were imported as response variables (Y) and the relationship between Y and the 82 previously generated VS+ descriptors (X) was evaluated with the PLS algorithm. Validation of the model utilised a number of different statistical techniques. BR analysis identified the major role played by hydrogen bonding donor solute properties in governing Caco-2 permeability.