Genetic Drift

Genetic drift is known as a very important factor when studying evolutionary systems. Drift is the opposite of selection, and makes sure the evolutionary process always keeps exploring its search space. In specific cases, genetic drift can be the leading force behind the dynamics of an evolutionary system.

What? An example

Imagine a cave filled with spiders and other animals. For some reason, the cave can become disconnected from the rest of the world, and the evolution of the animals within the cave will go on without any influence from what's happening outside of the cave.

Because of the darkness within the cave, there's no use for having eyes. Consequently, there will be no selective pressure on the genes that determine the phenotype of the individual's eyes, and mutations will become the driving factor behind the evolution of the genes responsible for the eyes. The composition of the genes for the eyes of the spiders and bugs in the cave's population will start to drift away, and the animals will lose their ability to see.

Combination of genetic drift and selection based on fitness

(TODO)

The importance of finite populations

Many models that study the dynamics of evolutionary systems assume populations with an infinite number of individuals. In such systems, the most probable population of the next generation can be computed easily, and the progress of the evolutionary system can be studied deterministically. In such a setting, the influences of genetic drift are completely discarded, as it is hypothesized that these influences are countered by the selective progress on the long run anyway.

Once these models and their predictions are compared to the predictions of stochastic, finite population models, the importance of genetic drift becomes clear. Mutations, and explorative search of the evolutionary system cannot be simply discarded in populations with a small population size. Even worse, in these small populations the importance of genetic drift can induce completely different characteristics when studying the long run behavior of the systems. As an example, if you're studying evolutionary game theory, the characteristics and existence of the fixed points of a inifinite population models can change dramatically. I'm currently studying the influence of finite population size on these changes in the fixed points.