Resistance is not futile

Our paper is out now, so I can finally explain what it's all about. Unfortunately I haven't had much time between grading tests, speaking to journalists and preparing for a talk tomorrow. It remains to be seen how much noise this will make, but I'm pleased with the initial response. Here are some clippings for you:

• Sex: why bother? (UH Press Release)
• How sex stays in fashion (Nature, Editor's Summary)
• Why we have sex: it's cleansing (Live Science)
• Sex: why even educated fleas do it (Ciencia Hoje)
  
• Sexo tem vantagens evolutivas (Público)

If you've read the earlier posts, you're probably wondering what robustness could possibly have to do with the evolution of sex? At first it would seem that positive epistasis, where mutations spontaneous mutations have weaker effects in combination than they do separately, is a "many-mutation" version of robustness. However, in the last few years, scientists working on a range of models of biological systems (such as, RNA folding and virus growth) have found that high robustness to mutation (that is, to one mutation) tends to be associated with more negative (or less positive) epistasis. In other words, it seems that selection for high robustness might lead to the evolution of negative epistasis.

So how does robustness to mutation evolve? The obvious answer is that high mutation rates can select for robustness to mutation. This has been known for a while. A less obvious suggestion is that sex might act as a qualitatively similar perturbation to mutation and lead to the evolution of robustness to mutation. To our surprise when we started working on this problem, although this had been predicted (most clearly in an uncharacteristically obscure paper published by Stephen Stearns in a Polish paleontological journal), it had never actually been tested in real organisms or even modelled in any detail.

To explore these questions we took a simple model of a genetic system from the literature. Briefly, it represents a very simple transcriptional regulation gene network. We then looked at the evolution of a very simple phenotype: the ability to produce a stable gene expression pattern. Population of these digital organisms were allowed to evolve such that those that produced stable gene expression patterns survived, while those that produced unstable patterns died. We found that sexually reproducing populations evolved increased robustness to mutations when compared to asexual ones. In addition, negative epistasis evolved in the sexuals, but not the asexuals. In other words, sexual reproduction created the conditions that favor its own perpetuation. If these results hold for real organisms, sex may be self-reinforcing.

Update: Dan Jones at The Proper Study of Mankind has written an excellent post on our paper.