Age-related macular degeneration (AMD) is the most common cause of blindness in developed countries and as such represents a very significant public health problem a number of specific genes, and the discovery, characterization, and eventual therapeutic control of these genes represent major goals of the vision research community. Although the strategies for gene discovery have become very powerful in recent years, there remains a major obstacle to the discovery of genes that underlie common, late-onset diseases like AMD. That obstacle is that clinicians cannot reliably sort patients with different molecular subtypes of late-onset disease into sufficiently homogeneous groups. The purpose of this project is to use the power of multi-agent systems computer technology in a novel way to aid clinicians in the collaborative development of a robust classification system based upon the ophthalmoscopic features of AMD. The result of this project will contribute to an NIH's Innovations in Biomedical Information and Science and Technology Program goal of speeding the progress of biomedical research through the development tools for electronic collaboration that will have impact on broader areas of biomedical research. ? ? We hypothesize that a multi-agent approach to this problem will result in a classification system with greater reproducibility and discriminative power than a system developed by clinicians without such computer assistance. The availability of populations of AMD patients with lower molecular complexity will significantly increase the power of statistical techniques for AMD gene discovery. In addition to this immediate and specific benefit, the strategies we will develop during this project for objectively interfacing medical experts with each other as well as with computers will have applications in the search for other late-onset disease genes as well as in the development of multi-center and multidisciplinary clinical trials of new therapeutic approaches. The proposed system, the Intelligent Distributed Ontology Consensus system (IDOCS) goes beyond conventional groupware by addressing drawbacks to direct, synchronous interaction by providing an autonomously coordinated, asynchronous interaction and collaboration platform among clinicians through their representative intelligent agents. IDOCS will provide a generic meta-data infrastructure using XMLJRDF to make it easily configurable for other diseases.
