Natural selection is often much like Goldilocks - an organism's traits shouldn't be too hot or too cold; natural selection likes them just right. In other words, traits are under pressure to remain near an optimum. If they deviate too far, natural selection will not-so-gently prod things back to the center. This phenomenon is known as stabilizing selection.
Stabilizing selection has to push against another powerful evolutionary force - random drift. Much of our genetic makeup is influenced by non-adaptive processes, that is, processes that are not particularly favored or disfavored by natural selection, and which do not perform some function that improves the fitness of the organism. Selection and drift have been especially hard to tease apart when it comes to gene regulation. Related species regulate their genes in different ways, but how many of those differences are simply due to random divergence? Trevor Bedford and Daniel Hartl at Harvard University take a crack at this question in a recent paper. They use a mathematical model based on Brownian motion (the kind of random motion you see when you watch pollen grains buffeted about in a drop of water) to determine how well stabilizing selection counteracts the battering of random drift.
It's clear that some genes are under very strong pressure to remain at an optimal level of regulation. Ribosome genes, which code for the critical components involved in translating the genetic message into protein, are uniformly expressed at high levels in all species examined so far. If your ribosome genes slack off, chances are you're going to lose the fitness game.
In other cases, it is not so clear how powerful stabilizing selection is. In their mathematical model, Bedford and Hartl treat random mutations like the tiny molecular collisions that knock around a pollen grain in Brownian motion. Natural selection is incorporated into the model by adding a slight twist to the idea of Brownian motion. In purely random motion, collisions (or mutations) do not push the system (a pollen grain, or a population or organisms) systematically in one direction. But in the presence of stabilizing selection, collisions that move the system towards an optimum are more favored than collisions directed away from the optimum.
Bedford and Hartl applied their model to gene regulation data from seven species of fruit flies. Physiologically, these species are fairly similar, but they have been diverging from each other for about 50 million years. Because of their similar physiology, we expect that many genes have been under stabilizing selection during this time. Random drift has played a role as well.
The researchers made two interesting discoveries. First, there were clear signs in the data that stabilizing selection had put the brakes on random drift; patterns of gene expression in these species were not as different as they would have been without stabilizing selection. Second, selection has been remarkably sharp-eyed. They found that small mutational changes in gene expression were unlikely to have a very significant fitness effect on the flies, but selection nonetheless was able to act on nearly imperceptible changes.
Bedford and Hartl note that this brings home the rather blunt statement by Alfred Russel Wallace: natural selection "has such enormous selecting power because of the overwhelming odds against the less fit." Given enough generations, the presence of natural selection becomes evident, even in subtle changes.
Join me tomorrow, here at Adaptive Complexity, for day 13 of 30 Days of Evolution Blogging Evolution as a science is alive and well. Each day I will blog about a paper related to evolution published in 2009.
Are you a blogger and want to join in? Here's how - and follow the link to check out how you can win a complete three volume set of The Life and Letters of Charles Darwin which includes an autobiographical chapter from Sir Charles and was edited by his son Francis Darwin, printed in 1887.