Research On Evolution, Functional Essentiality And Dynamic Modular Organization Of Proteins Based On The Integration Of Gene Expression And Protein Interaction Data

Protein-protein interactions play a critical role in many cellular processes, and networks of protein-protein interactions provide a systematic view of how the cell is functioning. One of the most important problems in systems biology is to find relations between a protein's topological properties and its functional features. It is found that proteins with higher degree evolve more slowly than proteins with lower degree. It is also found that proteins with higher degree are more essential than proteins with lower degree. One of the most challenging problems in systems biology is to study how proteins are organized to perform their cellular functions. Proteins are found to be dynamically organized into modules in the cell, which can help us better understand the dynamic properties of protein interactions. However, these findings have been heavily debated. In this work, we mainly focus on studying evolution, functional essentiality and dynamic modular organization of yeast proteins by integrating gene expression and protein interaction data.First, we integrate gene expression and protein interaction network data to derive a co-expressed protein-protein interaction degree (ePPID) measure, which reflects the number of partners with which a protein can permanently interact. We find that ePPID is a better predictor of protein evolutionary rate than protein-protein interaction degree (PPID), and the contribution of ePPID to protein evolutionary rate is statistically independent of other genomic variables. Analysis of hub proteins in the Structural Interaction Network further supported ePPID as a better predictor of protein evolutionary rate than the number of distinct binding interfaces, and clarified the slower evolution of co-expressed multi-interface hub proteins over that of other hub proteins. Our study demonstrates that ePPID can help us better understand how protein interactions affect protein evolutionary rate. Thus, we have at least in part resolved the longstanding debates on the relationship between PPID and protein evolutionary rate.Second, we integrate gene expression and protein complex data to propose a complex co-expressed density (CCD) measure, which is defined as the percentage of interactions within a complex that show significant co-expression. Our study shows that CCD can help better explain the relationship between protein complexes and co-expression. Furthermore, we find that co-expressed stable modularity is an important constraint that not only keeps proteins within complexes evolving at lower rates but also keeps protein pairs within complexes evolving at more similar rates. On the other hand, non-co-expressed transient modularity does not have such a property. Thus, we demonstrate that co-expressed stable modularity rather than non-co-expressed transient modularity exerts higher selective constrains on protein evolution.Third, we integrate gene expression, functional module and protein interaction network data to classify hubs into four types:non-co-expressed non-co-cluster hubs, non-co-expressed co-cluster hubs, co-expressed non-co-cluster hubs and co-expressed co-cluster hubs. The four hub types not only have different topological properties and functional consistence, but also have different enrichments in essential proteins. Furthermore, we find that the majority of hubs are essential due to their local role as core components of functional modules or protein complexes, but not due to their global role in organizing the whole network. Thus, we demonstrate that our classification of hubs into four types can better improve the understanding of gene essentiality, which will further help us better resolve the longstanding debates on the relationship between PPID and gene essentiality.Fourth, we integrate gene expression, protein complex and protein interaction network data to classify hubs into four types:non-co-expressed non-co-complex hubs, non-co-expressed co-complex hubs, co-expressed non-co-complex hubs and co-expressed co-complex hubs. From different aspects to analyze, the four hub types play different roles in organizing the network. Furthermore, the four hub types have different functional consistence. These results support the dynamically modular organization of the protein interaction network. Analysis of hubs in the Structural Interaction Network further validates the rationality of this classification method. Our study also shows that this classification method and the hub classification method in the previous paragraph have good correspondence and complementarity. This classification method can help us better understand the dynamically modular organization of the protein interaction network. Thus, we have at least in part resolved the longstanding debates on whether date and party hubs indeed exist in the yeast interactome.In summary, our study demonstrates that the integration of gene expression and protein interaction data can help us better mine the biological features which are implied in topological properties of protein interaction networks, and help us better understand how the cell is functioning from a systematic view.