Abstracting Complex Interaction Networks

The exploration of complex interaction networks has attracted considerable interest in various fields, ranging from fundamental biology and medicine to statistical physics and information technology.  In -omics disciplines, significant progresses have been made in understanding the large-scale properties and the biological relevance of these interactions. Some properties such as scale-free distribution of nodes connectivity or centrality are aspects commonly described in such complex interaction systems. In many of these studies the analysis of network topology is complemented by a semantic analysis that may rely on  different labels associated to the interacting entities.   One of the bottleneck of these semantic analysis is that they  are computationally costly. In this paper we present a framework to explore abstraction of networks useful to speedup the computation  of ground network measures. Such abstraction mechanisms may be used to efficiently provide accurate approximations of ground network measures.