Expert systems to aid in the formulation of hypotheses and design of experiments in biomedical research

Abstract

The recent National Research Council call for a unified biology-biomedical sciences research data base should also be taken as a challenge to develop an expert system to aid in its full utilization as a tool. While the sheer size of the body of knowledge in the biological and medical sciences is enough to warrant investment in a unified database, it is the implicit relationships across traditional research fields which hold the promise of rich returns. Additionally, access to facts or even to related clusters of facts does not make full use of the knowledge contained in the “biological matrix”. The ability to fully utilize that knowledge resides largely in the “art” portion of research and in less fuzzy, but field specific, heuristics. We describe the research process, including formulation of hypotheses, decomposition into subhypotheses, design of experiments, and use of results of those experiments. We also generally outline an expert system designed to aid in this process and discuss some specific problems encountered in its implementation.