The current Human Genome Project is largely devoted to structural characterisation of our genome. We now need international co-ordination of a second phase of genome analysis, the systematic construction of expression maps using both basic and high-resolution expression assays. Databases recording different types of expression pattern for a variety of human cell types need to be established and co-ordinated. There is a compelling need for a database of gene expression in early human development, but the scarcity of human material for study requires optimisation of research strategies and co-ordination of expression studies in early human and mouse development.