Ken A. Hawick
Computational biology, regulatory networks, betweenness centrality, complex network
Computational graph analysis metrics can be used to make quantitative comparisons between biological and regulatory networks obtained from real specimens with simulated synthetic networks that have well parameterized properties such as a scale-free structure. We compute the betweenness centrality metric for five public domain protein-protein network data sets and compare them with synthesised NK networks. We employ a node-culling procedure to progressively remove the highest connected nodes in these networks and assess the wholistic system changes as revealed by the resulting Floyd all-pairs distance and the number of component clusters in the networks as they fail and break up. We discuss the potential for this method in assigning characteristic signatures or categories to networks of this sort as well as for identifying network components that are most vulnerable to biological attack.
Important Links:
Go Back
