Classification models for CaCo-2 permeability using chemical information and machine learning techniques: Scope and limitations

The Caco-2 cell monolayer has been used across the pharmaceutical industry as an in vitro model of the human small intestinal mucosa to predict the absorption of orally administered drugs and a high correlation between the in vitro apparent permeability across Caco-2 monolayers and the in vivo fraction absorbed of drug candidates is established. In the present study, we have studied the role of different chemical information types and machine learning techniques on the predictive power of the resultant classification models. The models described has potential to be employed in a drug discovery environment as it is robust and covers a large chemical space.