KNOWLEDGE BASED BIOLOGICAL MODELING INFORMATION SYSTEM

生物医学研究模型委员会建议， 一个“生物知识矩阵”（矩阵），以解决需要 一种组织和智能访问财富的方法， 生物医学研究人员面临的信息（c.f.模型 Biomedical Research，National Academy Press，1985）。 到 实现矩阵将需要融合人工 智能（AI）和数据库技术以及 了解生物学知识的结构。 然而，在这方面， 最好不要尝试全面实施矩阵， 对有限知识领域的试点研究， 证明了这一概念的可行性和实用性。 一 相当普遍感兴趣的领域是核酸领域， 蛋白质序列库，这已成为宝贵的研究， 为生物医学研究界的很大一部分提供工具。 除了图书馆本身之外， 已经编写了程序来分析序列数据， 带来了重大的新发现 这些的效用 图书馆的典型实验室研究人员可以大大 通过增加他们的知识类型， 管理和推理系统设想的矩阵。 这项工作的目标是建立一个矩阵的测试平台 在现有序列库之上的概念， 矩阵的更大目标和 序列库用户。 这一进程的步骤将包括： (1)以关系形式实现GenBank数据库， 与人工智能环境的接口;（2）创建一系列 与数据库相关的知识库，从简单的 分类法和向更复杂表示法的发展（在 较窄的测试域）;（3）创建一组用户界面 以分子生物学家的需要为导向;（4）调查 对知识库进行推理的各种技术， 通过类比进行特定推理，最终目的是允许 用户可以在本地建立和测试假设。

The Committee on Models for Biomedical Research has proposed a "matrix of biological knowledge" (Matrix) to address the need for a means to organize and intelligently access the wealth of information that confronts biomedical researchers (c.f. Models for Biomedical Research, National Academy Press, 1985). To implement the matrix will require a fusion of artificial intelligence (AI) and database technologies as well as an understanding of the structure of biological knowledge. However, a full implementation of the Matrix is best not attempted until pilot studies on a restricted domain of knowledge have demonstrated the feasibility and utility of the concept. One domain of considerable general interest is that of nucleic acid and protein sequence libraries, which have become invaluable research tools for a large segment of the biomedical research community. In addition to the libraries themselves a large number of computer programs have been written to analyze the sequence data, which has resulted in major new discoveries. The utility of these libraries to the typical laboratory investigator could be greatly enhanced by augmenting them with the types of knowledge management and reasoning systems envisioned for the Matrix. The goal of this proposed work is to build a testbed for the Matrix concept on top of the existing sequence libraries, benefitting both the larger goals of the Matrix and the immediate needs of sequence library users. Steps in this process will include: (1) implementation of the GenBank database in a relational form, interfaced to an AI environment; (2) creation of a series of knowledge bases relating to the database, starting with simple taxonomies and progression to more complex representations (on narrower test domains); (3) creation of a set of user interfaces oriented to the needs of molecular biologist; and (4) investigation of various techniques of inference against the knowledge base, in particular reasoning by analogy, with the ultimate aim of allowing users to build and test hypotheses locally.