L. F. Jardine, A. O. Krassavine, A. W.R. Payne and S. Porter Pages 87 - 93 ( 7 )
In order to increase the rate of drug discovery, pharmaceutical and biotechnology companies spend billions of dollars a year assembling research databases. Current trends still indicate a falling rate in the discovery of New Molecular Entities (NMEs). It is widely accepted that the data need to be integrated in order for it to add value. The degree to which this must be achieved is often misunderstood. The true goal of data integration must be to provide accessible knowledge. If knowledge cannot be gained from these data, then it will invalidate the business case for gathering it. Current data integration solutions focus on the initial task of integrating the actual data and to some extent, also address the need to allow users to access integrated information. Typically the search tools that are provided are either restrictive forms or free text based. While useful, neither of these solutions is suitable for providing full coverage of large numbers of integrated structured data sources. One solution to this accessibility problem is to present the integrated data in a collated manner that allows users to browse and explore it and also perform complex ad-hoc searches on it within a scientific context and without the need for advanced Information Technology (IT) skills. Additionally, the solution should be maintainable by in-house administrators rather than requiring expensive consultancy. This paper examines the background to this problem, investigates the requirements for effective exploitation of corporate data and presents a novel effective solution.
blockbuster model, independent software vendors (ISVs), Uniform Resource Names (URN) system, QueryConstructor(tm), decision-making
Intellidos Ltd., HarwellInnovation Centre, 173 Curie Avenue, Harwell, Didcot, Oxford Shire, OX11OQG, UK.