A model-driven approach to industrializing discovery processes in pharmaceutical research

Despite year-to-year increases in R & D budgets, the number of successful NCEs (new chemical entities) has continued to decline. Drug companies are looking into new ways to make research processes more efficient, to manage information better, and to improve collaboration among research groups. This shift from an artisan approach to an organized, streamlined discovery process is often termed the "industrialization of discovery processes." This paper presents an approach to industrializing drug discovery that involves the formal modeling of research processes at several layers of abstraction, mappings between adjacent layers, and an implementation of this hierarchy using information technology-level execution elements. This approach is applied to the assay development phase of the drug discovery process. First, a business operations model is built by identifying the business artifacts, developing models for the life cycles of these artifacts, and then creating a comprehensive model that combines these life-cycle models and their interactions. Using the concept of adaptive business objects, a solution composition model that expands on the business operations model is developed. This model is then mapped into an executable platform-specific implementation by using the IBM WebSphere((R)) platform. Our prototype system was built and validated as a joint effort between IBM Research and Bayer HealthCare.