‘Better Compounds Faster’ – The Development and Exploitation of a Desktop Predictive Chemistry Toolkit

Posted by

SELECTBIO

Today’s drug designer has access to a vast quantity of data and a huge number of sophisticated computational methods. At the same time there is increasing pressure on the pharmaceutical industry to improve its productivity and reduce candidate drug attrition. In AstraZeneca we recognised both the need and the opportunity to build a strong Predictive Chemistry capability comprising an efficient and well-defined design process supported by highly integrated design and knowledge tools underpinned by the best predictive chemistry models. Here we describe the evolution of the AstraZeneca Predictive Chemistry integrated desktop toolkit and its exploitation by design teams to make better design decisions during the lead identification and lead optimisation phases. The toolkit comprises a mixture of in-house developed and commercial applications for data visualisation, Structure-Activity Relationship (SAR) analysis, virtual compound generation, property prediction and target compound selection, linked to a collaborative annotation and knowledge capture environment. The guiding principles for the development of the toolkit have been (a) to facilitate diverse design and analysis workflows in a flexible manner, (b) to use the best available validated computational and cheminformatic methods, (c) to develop user-friendly, robust and supportable desktop applications. The exploitation of the toolkit is driven through extensive Awareness, Training and Utilisation activities including web portals, on-line guides, help pages and video demos, classroom hands-on training using real project examples and ‘hot-desking’ by informaticians with design teams. We will describe some of the challenges we have encountered along the way, both technical and cultural, and highlight the benefits of close team working between informatics, medicinal and computational chemistry functions.