An architecture for automated reasoning systems for genome-wide studies

The massive amounts of data generated by high-throughput experiments makes modern biomedical research a data-intensive discipline, shifting the research methodology from a hypothesis-based approach to a hypothesis-free one. A formal procedure should be defined to properly design a study, understand the outcomes and plan improvements for each task performed during the experiments. Such formal approach needs the identification of a high-level conceptual model of the knowledge discovery process occurring in genome-wide studies: this is what existing computational tools lack. Starting from an epistemological model of the discovery process proposed for diagnostic reasoning, we describe how the design and execution of modern genome-wide studies can be modelled using the same framework. We show the general validity of the model, how it can be instantiated to model typical scenarios of genome-wide studies, and how we use it to develop tools aimed at building semi-automated reasoning systems.