Functional topology in a network of protein interactions

Natasa Przulj1, Dennis Wigle2, Igor Jurisica
1natasha@cs.toronto.edu, Department of Computer Science, University of Toronto, Canada; 2, Department of Surgery, University of Toronto, Canada

Protein-protein interactions (PPI) are building blocks of biological networks. The largest currently available PPI dataset is the S. cerevisiae cumulative PPI network consisting of over 78,000 interactions between 5,321 proteins. We represented the PPI network as a graph, where nodes correspond to proteins and edges connecting them are interactions. We analyzed the network using diverse graph theory approaches to obtain insight into the inner workings of cells and determine the structure-function relationships. We integrated the results of this analysis with existing functional annotation databases to construct computational models for describing and predicting the properties of lethal and viable mutations and proteins participating in genetic interactions, functional groups, protein complexes, and signaling pathways. Our analysis suggests a distinct property of lethal mutations: they are not only highly connected within the network, but also their removal causes a disruption in network structure. In addition, we provide evidence for the existence of alternate paths that bypass viable proteins in PPI networks, while such paths do not exist for lethal mutations. We also examined different functional classes of proteins and showed that they have differing network properties. We demonstrated a way to extract and predict protein complexes and signaling pathways. We have validated our predictions by comparing them to a random model, and the accuracy of predictions is assessed by analyzing their overlap with MIPS database. The models that we developed provide a means for understanding the complex wiring underlying cellular function, and enable us to predict essentiality, genetic interaction, function, protein complexes and cellular pathways. Results of our systematic analysis suggest that there is a relationship between network structure and protein function.