The ability to capture and access structured experimental data is a prerequisite to robust data analytics. Measured compound attributes such as biorelevant solubility and formulation characteristics are key components of predictive models for oral absorption. Leveraging these models in discovery can guide molecule selection and formulation development prior to preclinical in vivo studies, and help assess bioperformance risk for compound prioritization purposes. A platform was built within electronic laboratory notebooks for the structured data capture of properties relevant to oral absorption, and the generation of contextualized datasets. The platform eases access to these data and merging with a large preclinical pharmacokinetics database, thereby enabling analysis and visualization. Importantly, predictive tools for preclinical bioperformance were built and implemented. These include relatively simple bioavailability predictors based on a small number of compound attributes, as well as more complex models - built using experimental datasets - that relate formulation selection and predicted PK.
Learning Objectives:
Generate more value from large volume of experimental physicochemical results impacting oral absorption, by using a structure data capture approach and strengthening visualization and analysis capabilities.
Guide compound optimization and selection in the Discovery space and progression through early development by enabling structure-property relationship from large, contextualized datasets combining in vitro and in vivo results
Design and implement predictive tools for preclinical bioperformance including simple bioavailability predictors as well as more complex models that relate formulation selection and predicted PK.
