The amount of publicly available experimental data in databases such as UniProt, PDB, ChEMBL, and PubChem is rapidly increasing. However the vast number of possible protein-small molecule combinations is still far from covered. Therefore effective eScience technologies are needed for the integration of this large amount of heterogeneous data to enable development of predictive models of drug selectivity, polypharmacology, or off-target effects. The 3D-e-Chem project aims at the integration of large amounts of heterogeneous data from various sources, development of protein structure-based chemogenomics tools by the creation of the 3D-e-Chem-VM Virtual Machine with existing and novel cheminformatics tools (KNIME nodes) and chemogenomics workflows.