Predicting the absorption of new chemical entities in early drug discovery is crucial, as it facilitates identification of compounds with suboptimal absorption properties, caused by, e.g., intestinal precipitation. However, existing prediction tools often do not provide a quantitative relationship between precipitation measured in vitro and its actual impact on the absorption of weak bases in humans. To address this shortcoming, we developed an innovative oral absorption prediction tool to investigate the relationship between precipitation and oral absorption. The tool utilizes easily accessible in vitro solubility, precipitation, and permeability data to build a decision tree based on the Classification and Regression Trees machine learning algorithm, ultimately allowing for comprehensively evaluating the role of precipitation vs. solubility for oral drug absorption. The decision tree is tailored for application during drug discovery, enabling early ranking and selection of drug candidates based on their absorption characteristics, while also providing flexibility for various research needs.
Learning Objectives:
Upon completion, participant will be able to describe how to process easily obtainable in vitro data using an algorithm-based decision tree for predicting oral absorption of basic drugs in drug discovery and understand the principles and application of this novel prediction tool.
Upon completion, participant will be able to recognize the critical role of in vitro precipitation for basic drugs during the early stages of oral absorption predictions, driven by the gastric-to-intestinal pH shift that leads to supersaturation and precipitation.
Upon completion, participant will be able to outline the challenges associated with the development of an absorption prediction tool.
