Thursday, April 09, 2009

New Breakthrough in Evolution Theory

Imagine a culture of yeast cells growing in a medium where sucrose is the only carbon source. Sucrose isn't the preferred carbon source for yeast but yeast can handle it if need be. Cells secrete an enzyme called invertase that breaks down sucrose to glucose and fructose

The products of the reaction, glucose and fructose, can be taken up by the cell or they may diffuse away before being taken up. Molecules produced by the invertase from one cell can be absorbed by a neighboring cell.

As the concentration of free glucose and fructose rises in the medium, cells that lack the ability to synthesize and secrete invertase may survive. Thus invertase negative mutants may accumulate because they don't need to make their own invertase in order to have a source of carbon. In game theory, such mutants are called "cheaters."

Hands up, all you people who think that the existence of a stable equilibrium of cheaters and cooperators is a new discovery in evolutionary theory.

Right, it's not.

The editors of Nature think it is, so they published the paper from the Dept. of Physics at MIT (Gore et al., 2009).

The MIT PR department thought it was revolutionary enough to warrant a press release that was picked up by ScienceDaily [Cooperative Behavior Meshes With Evolutionary Theory].
One of the perplexing questions raised by evolutionary theory is how cooperative behavior, which benefits other members of a species at a cost to the individual, came to exist.

Cooperative behavior has puzzled biologists because if only the fittest survive, genes for a behavior that benefits everybody in a population should not last and cooperative behavior should die out, says Jeff Gore, a Pappalardo postdoctoral fellow in MIT's Department of Physics.

Gore is part of a team of MIT researchers that has used game theory to understand one solution yeast use to get around this problem. The team's findings, published in the April 6 online edition of Nature, indicate that if an individual can benefit even slightly by cooperating, it can survive even when surrounded by individuals that don't cooperate.

In short, the study offers a concrete example of how cooperative behaviors can be compatible with evolutionary theory.
I agree that this is an interesting example but I don't think the public is well served by presenting it as a new contribution to evolutionary theory. The public is entitled to think that evolutionary biologists must be really stupid if they've never thought of this before.

They (the public) would be really confused if they happened to read the Wikipedia entry on John Maynard Smith (1920 - 2004).

There are no references to Maynard Smith's work in the citations at the end of the Nature paper, although there is a reference to "Smith, J.M." who wrote the book Evolution and the Theory of Games.

Gore, J., Youk, H., and van Oudenaarden A. (2009) Snowdrift game dynamics and facultative cheating in yeast. Nature advance online publication 6 April 2009. [DOI: doi:10.1038/nature07921]


  1. I too was mistyfied why anyone thought this work merits publication in Nature. My first thought was "so, what's the big deal?" OK, so yeast play snowdrift but not prisoner's dilemma. Isn't it somewhat obvious? E.g., a paper that is not cited, Behavioral Ecology (2006), 17(4):633-641, "Asynchronous snowdrift game with synergistic effect as a model of cooperation" says:

    Greig and Travisano (2004) suggest that yeast cells play the PD on a sucrose plate. Here, a fraction of the population is able to secrete an invertase enzyme that transforms sucrose to glucose, which can be stolen by neighboring cells that do not produce the enzyme due to the lack of the gene. In this case, it is evident that without the production of the invertase enzyme none of the cells can grow because they are unable to digest the only available resource. Thus, S > P and the yeast cells play a snowdrift game instead of a PD one.

  2. I didn't find this paper as bad as your description led me to expect. I agree that it's not an earth-shattering discovery of the sort that Nature likes to think it specializes in, but as a "concrete example" (to use the term used in the press release) of something previously known mainly from computer simulations it's not bad, and I'm not sure the authors claim much more. Even if their reference to "J. M. Smith" suggests a rather hazy familiarity with one of the great evolutionary biologists of the past half-century, they do seem to know about game theory and its terminology -- doves, hawks, cheaters etc.

    The lack of citation to the papers DK mentions may be more serious -- I need to look to see in more detail what they say.

  3. Shades of On the Evolution of Bacterial Chromosomes, New viral way of life discovered in deep-sea vents, etc. It seems that instead of Newton's "If I have seen further, it is by standing on the shoulders of giants." [and yes, I am fully aware that the roots of the aphorism lie further back in history], we now live in the era of "If a quick Google search doesn't find it, it is a new observation." *sigh* A colleague tells me it just means I'm getting old and crotchety.

  4. Either one of the reviewers was the mother of one of the authors, or some Nature editor lost a bet somewhere, because this paper appears to be only a very minor extension of this paper (reference #15):

    Greig, D. & Travisano, M. The Prisoner's Dilemma and polymorphism in yeast SUC genes. Proc. R. Soc. Lond. B 271 (suppl.). 25–26 (2004)

    The key difference seems to be that the 2004 paper used yeast grown on agar plates, while the 2009 paper used yeast grown in suspension.

  5. I do not know much about Game Theory but isn't there a simpler explanation for the observed phenomenon.

    It was written in the paper: "However, when the initial fraction of cooperators was low, we found that the frequency of cooperators increased, suggesting that in the steady state there will be coexistence between the two strains."

    If the initial fraction of cooperators is low, doesn't it mean not enough glucose is produced in the culture. As glucose concentration is always high near cooperators while can be low for some cheaters (those far away from cooperators), cheaters are dying (starving) at a higher rate than cooperators. Although cheaters near cooperators may be proliferating at a higher rate than cooperators, the gain is offset as many other cheaters are dying off.

    Hence, in the beginning, net gain of cooperator is actually greater than net gain of cheaters. However when cooperator increases, net gain of cheater increases (more cheaters are near cooperators now hence less starving). However, when the cooperators increase, death rate will also increase as well (less are near cooperators). Eventually, an "equilibrium" is reached. Equilibrium is quoted because there be should very small alternating increase/decrease between cooperators and cheaters much as observed in the nature between population of prey and predator.

    I just do not understand how the Snowdrift theory come in. The two yeast strains are not making any decision or there are feedback mechanisms that govern whether the cooperator want to cooperate or whether a cheater want to cheat?.

    The whole thing can be simply be modeled as dynamic interaction between death/proliferation rate and glucose concentration etc and need no game theory.

    I may come across as stupid and ignorance but please tell me if I am indeed so. Will not be offended.

  6. I retract my previous critique.

    On more careful reading, this is a beautiful paper and is destined to become a classic in the field. True, it doesn't have the novelty of result, or sexiness of field so craved by journals like Nature, but the controls are exquisite, particularly the ability to adjust the fitness of the cooperators by manipulating histidine levels, and the analysis is perceptive.

    Ordinarily I'm down on Nature for valuing flash and sizzle more than substance and precision, so I have to commend them for publishing this paper.

  7. It's in nature because it's Van Oudenaarden, possibly the brightest mind in the systems bio field, and he's from a physics department, and it's using yeast invertase, an enzyme most scientists know. The quality of the paper doesn't actually matter for acceptance into Nature.

    Also, the MIT press release doesn't really exaggerate anything--it doesn't mention other work on evolution of cooperation, but it makes the point that this is key since it provides a "concrete example" of such behavior, for which, on the molecular level, we only have a handful of examples.

  8. It's in nature because it's Van Oudenaarden, possibly the brightest mind in the systems bio field

    I looked him up on Pubmed. Wow. Absolutely amazing publication record with Science and Nature on almost annual basis. Maybe I am an idiot but I am totally underwhelmed by the content. Take, for example, Science (2008), 319:482-4, The Frequency Dependence of Osmo-Adaptation in Saccharomyces cerevisiae:

    So, what's the big deal there? That you can model any simple oscillation with feedback loops? This was in textbooks at least 20 years ago. That Hog1 is a major regulator of osmolarity sensing? That was established at least 10 years ago. That gene expression plays no role in responses on a 5 min scale? Uh, that's a really surprising conclusion (NOT). So, what else is there??? What new knowledge about living systems should I have gained from reading this paper? Oh, and this is the conclusion:

    "These results demonstrate the promise of applying engineering principles to cellular
    networks, particularly when predicting the
    response of the system to dynamic stimuli. In
    more complex systems, measuring the activity
    level of all relevant state-space variables could help with determining the effective network structure."

    No shit. Sigh.

    The quality of the paper doesn't actually matter for acceptance into Nature.

    Wonder if Nature editors agree :-)