Analysis of Biomedical Networks with BioNAR
Date:
Author(s): Anatoly Sorokin, Colin Mclean, J. Douglas Armstrong, Oksana Sorokina
Networks and graphs are ubiquitous in biology, they are used to represent a wide range of concepts from intermolecular interactions in the protein complexes, to gene-disease associations, and food networks in ecology. In the analysis of such networks, scientists would need to identify keystone nodes, find clusters of highly connected nodes, and annotate edges and nodes with numerical or categorical information. We are presenting the Bioconductor package BioNAR, which combines a selection of existing R protocols for network analysis with newly designed original methodological features to support step-by-step analysis of biological/biomedical networks. BioNAR provides functionality for network creation and annotation; calculation, storage and visualisation of basic graph properties and centrality measures; identification of network communities and evaluation of their robustness; comparison of the communities created by different algorithms with each other and with external node annotations; identification of keystone nodes; and inferring dynamics of the system represented by the network of interacting nodes. We illustrate BioNAR functionality with three cases describing (i) analysis of the synaptic proteome interaction architecture, (ii) identification of key genes in the ‘diseasome’ network, and (iii) prediction of steady-state abundances in the microbiome bacterial interaction network.