Beyond classical domains: The protein unit paradigm - from protein-architecture characterization to downstream...

The classical concept of a protein domain has impacted structural and molecular biology, but geometric, evolutionary and functional criteria often result in conflicting domain boundaries. We proposed the "protein unit" (PU), an intermediate, evolutionarily conserved structural motif that connects secondary structures to complete domains. A hierarchical decomposition of protein structures identifies these compact units and their hierarchical arrangement, clarifying ambiguous boundaries and yielding multiple plausible domain organizations. Considering proteins as assemblies of modular PUs opens new perspectives and applications, such as structural comparison, by accounting for relative movements, insertions and circular permutations, as well as improved template detection in comparative modelling. Structural unit libraries also enable a modular design strategy in which new architectures arise from alternative combinations of existing units. The protein-unit framework redefines protein architecture as a hierarchy of reusable modules and opens new avenues of research for prediction, modelling and protein engineering.