Investigating the structure of integrated biological networks

Theory of complex networks provides an intuitive setting for studying biological relationships at the cellular level and beyond in the topological context [1]. It has already been suggested that the topological properties of networks relate to biological function [2]. However, due to diversity of biological relationship types even at the molecular level, studying the biological network topology at one level only (e.g. metabolic networks) may miss much of important information about cross-talk across multiple pathways and potential feedback loops via regulatory networks. Our aim is to develop a framework to study topology of multiple biological networks. As a start, we mapped KEGG, TransFac, TransPath, MINT, and BIND databases using XML [3], with example shown in Figure1. We developed a Java-based tool that allows parallel retrieval across multiple databases, incl. metabolic pathways, protein-protein interactions, signalling and regulatory networks. The results are then visually displayed as a network (Figure2). Edge attributes contain information about type of relationship, possibly quantitative or semantic information (such as “is located in” in case of linking a protein with a complex entity such as cell organelle). Information in integrated network form is a starting point for deeper topological and functional mining. Presently, we are primarily interested in the concept of similarity, i.e. how to define a distance metric across multiple networks. We applied various distance metrics and nonlinear mappings into lower-dimensional space such as Sammon’s mapping and self-organizing maps, of which results will be presented. We are extending our approach to more complex entities by combining existing pathways with the automated text mining. We believe our approach will provide a powerful framework for context-based mining and modelling of biological systems.