A multiple-species, multiple-project database for genotypes at codominant loci

The role of a genotypic database is identified as an interface between raw data sources, e.g. genotyping laboratories, and the ultimate use of the dat a in research or marker-assisted selection. Structure and features of ADRDB , a relational database for codominant marker loci, are described. Various potential reasons for data inconsistencies, such as null alleles, mutations or false paternity, are reviewed. A detailed description of all tables is followed by an explanation of the algorithm used for error detection and, w here possible, derivation of unknown genotypes from relatives. The ADRDB cu rrently contains roughly 300 000 genotypes from three different species and is integrated with programs and performance data for QTL research. Though initially developed for scientific purposes it may be seen as a prototype f or a genotypic database in livestock breeding programmes.