Integrative analysis of human protein, function and disease networks.

Wei Liu,Aiping Wu,Xiaofan Wang,Matteo Pellegrini

doi:10.1038/srep14344

Abstract

Protein-protein interaction (PPI) networks serve as a powerful tool for unraveling protein functions, disease-gene and disease-disease associations. However, a direct strategy for integrating protein interaction, protein function and diseases is still absent. Moreover, the interrelated relationships among these three levels are poorly understood. Here we present a novel systematic method to integrate protein interaction, function, and disease networks. We first identified topological modules in human protein interaction data using the network topological algorithm (NeTA) we previously developed. The resulting modules were then associated with functional terms using Gene Ontology to obtain functional modules. Finally, disease modules were constructed by associating the modules with OMIM and GWAS. We found that most topological modules have cohesive structure, significant pathway annotations and good modularity. Most functional modules (70.6%) fully cover corresponding topological modules, and most disease modules (88.5%) are fully covered by the corresponding functional modules. Furthermore, we identified several protein modules of interest that we describe in detail, which demonstrate the power of our integrative approach. This approach allows us to link genes, and pathways with their corresponding disorders, which may ultimately help us to improve the prevention, diagnosis and treatment of disease.

Highlights

Disease Modules h t k jghtkjlghtkjl a protein module is defined as a group of proteins that carry out similar functions
Topological modules represent a locally dense structure in protein-protein interaction (PPI) networks; function modules represent the aggregation of proteins of related function in a function network; disease modules represents a group of proteins that share a common disease phenotype within a disease network
Starting with the protein interaction dataset from Hippie[29], we identified 136 topological modules with NeTA, 136 corresponding functional modules, and 139 disease modules annotated using OMIM31 and GWAS32

Summary

Introduction

Disease Modules h t k jghtkjlghtkjl a protein module is defined as a group of proteins that carry out similar functions. These functions are associated with the same pathway, and could be associated with a particular disease. Starting with the protein interaction dataset from Hippie[29], we identified 136 topological modules with NeTA, 136 corresponding functional modules (annotated using Gene Ontology30), and 139 disease modules annotated using OMIM31 and GWAS32. By systematically integrating the three levels of networks and protein modules, we found that our multi-level method for biological interpretation has distinct advantages over approaches that only consider subsets of data and annotations. The approach we present here provides an avenue for network integration, and promises to shed light on the prevention, diagnosis and treatment of complex diseases

Methods

Results

Conclusion