A Next-Generation Platform for Novel Drug Discovery Using Deep Learning of ADMET Properties.

Integrated through Optibrium's existing software offering (StarDrop(tm)), DeepADMET is a next-generation project that will develop and apply novel deep learning methods to extend and improve predictive models of important Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) endpoints to guide more efficient design and selection of high quality drug candidates. Data for important ADMET endpoints that are not readily accessible will be curated as the basis for building models for these endpoints, which are not currently available, and extend and improve the accuracy of the current state-of-the-art models. The successful delivery of project objectives will enhance Optibrium's foothold in the Computational Drug Discovery market through the commercialisation of the software licences through our existing sales force or, where appropriate, by partnership with organisations specialising in that sector or region. DeepADMET forms a key part of Optibrium's Augmented Chemistry(tm) strategy, moving beyond 'tools' to support chemistry optimisation. Intellegens is a spin-out from the University of Cambridge that has developed a unique Artificial Intelligence (AI) method for training neural networks from incomplete data sets. The MDC has a wide variety of partners drawn from: UK biotech and enables it to meet its core strategic objectives as a catalyst for enabling innovation within the sector. Specifically, the project utilizes key and rare expertise in information extraction and curation strategies and tools, the further development and validation planned here will be applicable in future translational life science sectors.