In recent years, after high throughput PPI data was available, studies have focused on unraveling how proteins organize their functionality from architecture of the PPI network. We examine the functional organization of a PPI network by dividing the network into layered structure around a protein according to shortest path length. We proposed an index, functional correlation, to assess the functional closeness of a specific protein with its l layer neighbors, i.e. proteins having l shortest path length from the center protein. Our results showed that functional correlation decays exponentially with the number of layers within a characteristic length l c , and it becomes uncorrelated outside such a characteristic length. A simple model based on functional unit structure was proposed to explain this exponential decay of functional correlation.
