It is known that enzymes can be composed of one or more, same or different, subunits. If some protein is composed of more then one subunit, it is called an oligomeric protein, 1 and one example of such proteins are purine nucleoside phosphorylases (PNPs). There are two main classes of this enzyme: trimeric PNP or “low molecular mass” protein which can mostly be found in eukaryotic organisms, and hexameric PNP or “high molecular mass” protein which can mostly be found in prokaryotic organisms (Figure 1).2 These two types od enzymes share only 20-30% sequence identity, but the overall fold of the single monomer is similar, and yet this similar monomeric building block makes a different quaternary structure. A large number of 3D structures of PNPs have been determined so far. We can base our understanding of the mechanism by which monomeric subunits communicate in these enzymes on this structural data, and take into account the oligomeric arrangement of subunits in enzymes from different organisms. As part of the ALOKOMP project (alokomp.irb.hr) which tries to understand the allosteric communication between these monomeric subunits, the relational database of PNPs is constructed. As a special subset of this database, an all-to-all overlap of these structures is also made, where we can cluster this data as well as compare some of the structures according to their assembly symmetry and their crystallographic symmetry.