Editors' ChoiceNETWORKS

Evolving Bacteria and Computers

+ See all authors and affiliations

Science Signaling  25 May 2010:
Vol. 3, Issue 123, pp. ec160
DOI: 10.1126/scisignal.3123ec160

As knowledge of biological regulatory networks has become more complete, it has been possible to begin to explore the engineering or design principles of such systems and to understand their evolutionary origin. To compare a biological system that we are beginning to understand to a similarly complex control system we fully understand because it was created by humans, Yan et al. compared the transcriptional regulatory network of the bacterium Escherichia coli to the regulatory architecture of the Linux computer operating system. The most obvious difference was in the hierarchical organization. The biological system has no network hubs (highly connected nodes) that are regulated by many other nodes in the network (no gene regulated by many different transcription factors) but does have nodes that regulate many others. The computer system has the opposite organization: lots of highly connected nodes that are regulated by many inputs and none that have a large number of outputs. Looking at regulatory modules (the set of nodes regulated downstream of one master regulator node), the authors saw much more reuse of nodes in the computer system, as well as much more overlap of modules. Randomly chosen modules in the Linux system overlapped by more than 80%, whereas in the E. coli network, average overlap was less than 5%. Interesting insights also came from comparison of the evolution of the two systems. Comparing 200 bacterial genomes and various versions of the Linux system, there were major differences in the conserved or persistent nodes and their network characteristics. In the transcriptional network, specialized nodes (used infrequently in other modules) were preserved, whereas in the computer system, there was a positive correlation between reuse and persistence. Programmers, of course, favor the reuse of modules, which provides a cost-effective way to build and develop a complex system. But the biological system cannot afford the cost of such a strategy in terms of robustness. If a frequently used module fails, it puts the whole system at risk, and, if changes are needed in a module, compensating changes may have to be redesigned into many other interacting partners—not so easy through natural selection in biology. Rather, the biological system has conserved components at the bottom of the regulatory hierarchy and seems to use changes in the regulators to drive evolution. But the authors point out that, in more complex genomes, reuse of components becomes more common (that is, the number of enzymes increases more slowly than the number of transcriptional regulators), perhaps more closely approximating the Linux system.

K.-K. Yan, G. Fang, N. Bhardwaj, R. P. Alexander, M. Gerstein, Comparing genomes to computer operating systems in terms of the topology and evolution of their regulatory control networks. Proc. Natl. Acad. Sci. U.S.A. 107, 9186–9191 (2010). [Abstract] [Full Text]

Related Content