Non-Adaptive Processes

September 24, 2007

Last week a review by Michael Lynch came out in Nature Reviews Genetics entitled, “The evolution of genetic networks by non-adaptive processes”.

In this review, Lynch takes a population genetics perspective on network evolution. Using relatively simple but powerful mathematical models of basic evolutionary processes, Lynch shows that many of the characteristic properties of networks (metabolic, developmental, and regulatory) that have been highly advertised lately (such as complexity, organization, robustness, evolvability, etc.) are just a basic consequence of the underlying nucleotide evolutionary process.

The key to evolutionary patterning of pathways is alpha = u_l / u_g, the ratio of site loss (u_l) to site gain (u_g) by mutational processes. Since a mutation which results in loss must occur within the regulatory motif u_l = nu where n is the site size (typically 4-12) and u is the per nucleotide mutation rate. For gain events, however, the mutation can occur at any place in the L nucleotides of the promoter (where promoter here is defined as locations where regulatory sites are capable of effecting change in the gene’s expression). Therefore u_g = (L-n+1)nu/4^n if we assume equal nucletoide frequencies.

From this simple kind of model, it is easy to show that a neutral process (ie. one without immediate fitness benefit) will still push the system towards higher complexity (ie. more binding sites and interactions). Soyer and Bonhoeffer (PNAS 103(44):16337-16342 2006) had a similar result using a chemical protein interaction model and doing simulations on all three protein interactions graphs. The interesting side effect of both of these works is that “robustness” may simply be a natural consequence … a byproduct of the system’s imbalance between size descreasing (u_l) and size increasing (u_g) mutations. While both of these paper’s models are simplifications of the underlying biology, it can easily be argued that the underlying principles they uncover will likely hold in more realistic models.

In the end, Lynch makes a solid argument for the need to include population genetics principles into subsequent computational models for network evolution. To approximately quote, “A mechanistic understanding of network evolution is unlikely to be achieved without specific references to ubiquitous non-adaptive forces of evolution that operate at the level of DNA (mutation and recombination) and populations (drift).”

This reminds me of Jonathan Eisen’s self-described obsession with adaptationism. In other words, existence is NOT evidence for selection, adaptation, robustness, or (insert your favorite current buzz-phrase here).

Lynch, M. (2007). The evolution of genetic networks by non-adaptive processes. Nature Reviews Genetics, 8(10), 803-813. DOI: 10.1038/nrg2192


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: