Contributions of network theory to biology

Metabolic cycles, gene expression, and other biochemical pathways are natural fodder for the network theorist. Like the internet, airlines’ route systems, and power grids, biochemical pathways contain branches and nodes. And like those manmade networks, biochemical pathways are complex, especially when you include their myriad regulatory checks and balances.

Despite that affinity, the application of network theory to biochemical pathways is relatively recent. Yesterday at the annual meeting of the Biophysical Society in Baltimore, Hawoong Jeong of the Korea Advanced Institute of Science and Technology noted that in 2000 only a handful of “network biology” papers appeared. Last year, he said, the number was around 1800.

Jeong made his observation at a session entitled “Contributions of Network Theory to Biology.” The first speaker and chair of the session was Sergei Maslov of Brookhaven National Laboratory. Maslov’s talk nicely exemplified one of those contributions: to make sense of the pathways’ daunting complexity.

Maslov and his collaborators look at, among other things, protein–protein interactions. In yeast cells, there are around 2000 different kinds of proteins. Those proteins interact with each other and with other molecules, including DNA and RNA.

Fully 80% of yeast proteins are connected in one giant network, whose scale-free topology resembles those of human social networks. Indeed, the median degree of separation of one yeast protein from any other protein in the network is the same as Kevin Bacon’s from any other actor : six.

The internet’s scale-free topology ensures myriad paths for information to flow from one node to another. But, noted Maslov, being scale-free is not obviously an unalloyed benefit to biochemical pathways. Unlike the internet, a biochemical pathway lacks the means to redirect traffic if one node malfunctions. Undesirable perturbations and interactions can spread.

One kind of interaction that may be harmless or undesirable is that between a protein and other proteins that don’t belong to its habitual pathway. In a cell containing N proteins, the rate of specific, pathway interactions is proportional to N, whereas the rate of nonspecific, off-pathway interactions is proportional to N².

In general, cells need N to be high enough to ensure their proteins can carry out their specific jobs efficiently but not so high that proteins waste too much of their time in nonspecific interactions. In a series of papers, the latest of which appeared last month, Maslov and his collaborators have applied network theory to analyze protein–protein interactions.

Among their conclusions: The abundance and concentration of proteins in cells and in two kinds of cellular compartment, mitochondria and nuclei, are more than high enough to ensure efficiency. In fact, according to Maslov and his colleagues’ calculations, proteins spend about 80% of their time interacting with proteins that belong to their pathways.

The remaining 20% of the time that proteins spend in the company of off-pathway partners might not be a complete waste. Just as Kevin Bacon and other humans benefit from meeting new people, nonspecific protein–protein interactions might promote evolutionary adaptation.

Charles Day