DIS '06 Proceedings of the IEEE Workshop on Distributed Intelligent Systems: Collective Intelligence and Its Applications
Efficient Detection of Network Motifs
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
| Hi-index | 0.00 |
“Network Motifs” is a research area in “Network Science” which has captured the attention of many researchers. Complex biological and social networks have displayed presence of some triad motifs far in excess (or short) of their expected values. Some of these over(under)-represented motifs have explained the basic functionality of systems, for example, in the sensory transcription networks of biology, over-represented motifs are shown to perform signal processing tasks. This suggests purposeful, selective retention of these motifs in the studied biological systems. Another interesting feature is the high correlation of triad motif significance profiles (MSPs) of all systems that belong to a family of naturally grouped systems, thereby suggesting that all the systems in a family have the same function to perform and hence the correlation. Engineering systems also display over(under)-represented motifs. The MSPs of a family of naturally grouped engineering systems show high correlation. Unlike biological and social networks, engineering systems are designed by humans and offer an opportunity for investigation based on known design rules. We show that over(under)-represented motifs in engineering systems are not purposefully retained/avoided to perform functions but are a natural consequence of design by decomposition. We also discover that naturally grouped systems have remarkably correlated in(out) degree distribution across nodes resulting in a high correlation in MSP. Therefore, we argue that the idea of network motifs has no significance in engineering systems (unlike biological and other evolutionary systems), and we caution the engineering research community to be careful while drawing conclusions based on network motifs. We report a remarkable correlation of in(out) degree distribution of systems within a family of engineering systems. We further show that biological and social networks also display signs of decomposition. They also show a high correlation of in–out degree distribution of nodes for systems within a family. This opens up an interesting opportunity to investigate these systems through their observed decomposition. © 2012 Wiley Periodicals, Inc. Complexity, 2012 © 2012 Wiley Periodicals, Inc.