Proteins are comprised of domains, which are conserved evolutionary units that often also correspond to functional units, and can frequently be detected with reasonable reliability using computational methods. Most proteins consist of two or more domains, giving rise to a variety of combinations of domain combinations. These combinations can be represented as graphs of domain associations and reveal intersting insights into the domain-wise evolution of proteins. We have developed a number of programs for the in depth analysis of domain-duplications, co-duplication of domains and circular permutations. Another level of complexity arises because proteins themselves can can form i complexes with small molecules, nucleic acids or other proteins. The networks of both domain combinations and protein interactions can again be conceptualised as graphs and these graphs can be analysed conveniently by computational methods. We summarise facts and hypotheses about the evolution of domains in multi-domain proteins and protein complexes, and the tools and data resources available to study them.
