Databases dedicated to molecular biology are an essential complement to literature data. Today there is a very wide variety of heterogeneous databases. This diversity is, of course, explained by the variety of biological data, which are not limited to sequences, but also by the variety of objectives which governed their design. The major problem in the management of biological data therefore does not result so much from their volume as from this heterogeneity, both in terms of nature and format. The fundamental question is therefore how to integrate these biological data in order to make them...