A private knowledge base for molecular biological research

The authors discuss an integrated knowledge-based (KB) system that is part of the molecular biological information processing system in the Japanese Fifth Generation Computer System (FGCS) project. The KB system is developed in the framework of a deductive object oriented database (DOOD) that consists of a knowledge representation language called Quixote and a database management system called Kappa-P. Kappa-P and Quixote are developed on a parallel inference machine (PIM). Kappa-P employs a nested relational model and has the extensibility suitable for parallel processing and amino acid/DNA sequence retrieval. Quixote provides sophisticated query processing capability with advanced concepts, such as objects with definite identity and subsumption relations, modules with submodule relations, and the flexibility to represent various forms of knowledge.