PAQS Data Sources, Exchange Formats, and Database Schemas for a Proteo-Chemometric Analysis and Query System

The research described in this Thesis is part of a project to develop a new database system for proteo-chemometric research. This new system uses a mediator/wrapper approach for integrating heterogeneous and autonomous data sources. Special-purpose modules for data representation and data analysis can be incorporated into the system through the extensibility of the object-relational mediator. Life science data sources and data exchange formats for the new Proteo-chemometric Analysis and Query System (PAQS) have been surveyed. Although important data sources exist on many different formats the trend towards XML is evident. For proteo-chemometric research it is important to be able to access data sources with binding affinity data. Most such data sources are only accessible via web forms, which limits the query capabilities. Database schemas for parts of the proteo-chemometric information domain have been developed within a functional data model with object-oriented extensions. These schemas have also been implemented in the Amos II system as a first-stage prototype of PAQS. Special emphasis has been put on modelling binding experiments and experiment evaluations, and the corresponding data types have been used to show how data analysis could be performed by means of foreign functions of the mediator. The mediator/wrapper approach is described in the Thesis, and examples are given of other systems which use this architecture for integrating life science data, both research prototypes and commercial systems. Introductions to the proteo-chemometric approach to drug design, to some general database concepts, and to information integration by means of database systems are also given.