Game Theory in Genetics
When it comes to natural selection, the phrase “survival of the fittest” generally means that the organisms that are best adapted to their respective environments will survive and pass on their genes to future generations. While this is mostly true, a recent study shows that genetic diversity may be more important than fitness. This discovery was made at the University of California, Berkley under the supervision of computer scientist Christos Papadimitriou. While Papadimitriou and his colleagues were studying a mathematical explanation for how sex is determined, they noticed similarities to an algorithm with linear programming and game theory applications.
This algorithm, known as the multiplicative weights update algorithm, will figure out how a player in a game should weigh his potential strategies when faced with a series of decisions. In the case of genetics, each gene will determine which of its alleles performed well and which performed poorly within the current environment. Then, using the multiplicative weights update algorithm, the weights of the positive performers will be increased and the weights of the negative performers will be decreased. As this genetic game gets played over and over again, it may help scientists to better understand why the fittest organisms don’t always drive their weaker competitors to extinction.
Although we did not specifically learn about the multiplicative weights update algorithm in lecture, its final product of genetic diversity draws some parallels to a different game: The Hawk-Dove Game. When the genes put weights on their alleles corresponding to their performance, the weights are not always assigned to make the fittest organism possible, as genetic diversity must be maintained. This is similar to how a player in the hawk-dove game would not want to choose hawk or dove every time because a dominant strategy in such a game is not mutually beneficial. Instead, a mixed strategy must be used in order to maximize a player’s payoff.
It would appear that the same thought process could be applied to this genetic game theory. If each gene assigned weights to its best-performing alleles, and thereby made the fittest organism possible every time, a genetically diverse population would be lost. A more homogenous population may not respond as well to a changing environment, and thus may be killed off. This is to say that if a gene chose “hawk” every time along with all the other players in the system, the expected payoff would be low as a homogeneous population is more susceptible to extinction. Consequently, if a gene chose “dove” every time, it would be missing out on creating more evolutionally fit organisms. So, by using a mixed strategy within the multiplicative weights update algorithm, an optimal payoff consisting of fit organisms as well as genetic diversity can be obtained.
Link:
http://www.simonsfoundation.org/quanta/20140618-the-game-theory-of-life/