Conception and development of a bibliographic database of blood nutrient fluxes across organs and tissues in ruminants: data gathering and management prior to meta-analysis.

In the organism, nutrient exchanges among tissues and organs are subject to numerous sources of physiological or nutritional variation, and the contribution of individual factors needs to be quantified before establishing general response laws. To achieve this, meta-analysis of data from publications is a useful tool. The objective of this work was to develop a bibliographic database of nutrient fluxes across organs and tissues of ruminant animals (Flora) under Access using the Merise method. The most important criteria for Flora were the ease to relate the various information, the exhaustivity and the accuracy of the data input, a complete description of the diets, taking into account the methods of the methodological procedures of measurement and analysis of blood nutrients and the traceability of the information. The conceptual data model was built in 6 parts. The first part describes the authors and source of publication, and the person in charge of data input. It clearly separates and identifies the experiments, the groups of animals and the treatments within a publication. The second part is concerned with feeds, diets and their chemical composition and nutritional value. The third and fourth parts describe the infusion of any substrates and the methods employed, respectively. The fifth part is devoted to the results of blood flows and nutrient fluxes. The sixth part gathers miscellaneous experimental information. All these parts are inter-connected. To model this database, the Merise method was utilised and 26 entities and 32 relationships were created. At the physical level, 93 tables were created, corresponding, for the majority, to entities and relationships of the data model. They were divided into reference tables (n = 65) and data tables (n = 28). Data processing was developed in Flora and included the control of the data, generic calculations of unknown data from given data, the automation of the estimation of the missing data or the chemical composition of the diets. It also included the construction of tables for meta-analyses and the study of the variations of several factors within publications (pre-coding of meta-analyses). Overall, the system was built to facilitate the gathering, input, validation, management and retrieval of data from publications.