Using JMP® for Data Analysis and Prediction in Drug Discovery.
The design and discovery process of new pharmaceutical agents is complex and expensive. In order to accelerate the process, saving both time and money, it is necessary to extract maximum knowledge from the vast quantity of data collected on millions of compounds at AstraZeneca. Using JMP we produce intuitive visualisations to better grasp trends and patterns in the data. We can use Design of Experiments to efficiently explore "chemical space." We use advanced modeling techniques to model the structure-activity relationships within our data and use these both to predict properties of future compounds and to guide compound design towards the desired properties. We are now able to predict advanced properties of compounds, such as an estimated dose to man, and – using the JMP profiler – simulate the effect of changes to multiple input variables in order to determine the design criteria for future compounds. Throughout, we use predictive techniques to guide compound design and enrich the properties of the synthesised compounds in order to reach the clinical candidate as rapidly as possible.