Workflow from Scientific Research

Overview of the data flow in this work. Dimer structures were predicted for proteins across four species proteomes, yielding 156,065 models. The models were subsequently scored based on a benchmark, yielding over 8,000 high-confidence dimer structures. The dimer structures were used to predict higher-order biologically relevant macromolecular assemblies into rings and filaments, yielding proteome-wide homo-oligomerization information.

DOI