AstraZeneca could substantially reduce animal studies by using Cerella to predict rat in vivo pharmacokinetic (PK) parameters.
The Challenge
To accurately predict in vivo PK parameters, such as exposure, clearance and bioavailability. scientists assess and predict both in vivo animal PK data as well as human and animal in vitro systems. This is an enormous challenge due to the many complex factors that can influence drug disposition and limited in vitro ADME data.
To improve design during drug discovery, help select compounds with better properties, and reduce the number of expensive, time-consuming in vivo experiments, AstraZeneca wanted to predict in vivo properties computationally – that’s where Cerella came in.
The Solution
By training Cerella models on AstraZeneca’s in vivo rat PK data for over 3,000 compounds, they were able to predict rat in vivo PK parameters with state-of-the-art accuracy based on a compound’s molecular structure and either measured or predicted in vitro parameters.
Cerella provided a fast and cost-efficient way to guide the design of molecules with optimal PK profiles, even when compared to other machine learning algorithms and deep learning approaches
It enabled the prediction of virtual compounds at the point of design, and supported the team in prioritising compounds for in vivo studies.