Cerella proved its value at a large scale, accurately predicting activity and ADME properties across Takeda Pharmaceutical’s data repository.
The Challenge
Working with drug discovery data is difficult, due to limited quality data, experimental errors and uncertainties, and the time and expense involved in collecting the data points needed
Takeda wanted to investigate how Cerella could support accurate predictions of activity and ADME properties on a large scale.
The Solution
Cerella’s deep learning methods were applied to Takeda Pharmaceutical’s data set comprising of approximately 700,000 compounds and 1,000 experimental endpoints.
A variety of applications were explored, including imputation and prediction of compound activities in project contexts, high-throughput screening results and a diverse range of ADME properties.
Amongst other things, they found that Cerella can make much more accurate predictions for the latest activity results, based on older data, and for a broad range of ADME properties, including complex cell-based results.