This article describes a novel deep learning neural network method and its application for the imputation of bioactivity data, such as assay pIC50 values. These deep learning methods underpin our Cerella technology, part of our Augmented Chemistry suite.

Summary

In this study, the team applied a deep learning neural network (the foundations of our Cerella platform) to impute assay pIC50 values. It was shown to outperform other approaches such as traditional QSAR models, and provide predictions with excellent accuracy.

Depiction of how our AI methods are able to support imputation of bioactivity data based on sparse assay data and molecular descriptors.

Citation details

T. M. Whitehead, B. W. J. Irwin, P. Hunt, M. D. Segall, G. J. Conduit, J. Chem. Inf. Model., 2019, 59(3) pp. 1197-1204. DOI: 10.1021/acs.jcim.8b00768

Find out more

Visit the journal webpage to read the full article, or some examples of Cerella in action by watching our webinars on predicting PK using limited ADME data or developing novel antimalarials.