**EvoWiki is now a project of the RationalMedia Foundation.We are moving all content to RationalWiki.See the EvoWiki project page for details!**

# Genetic drift

### From EvoWiki

**Theory of Evolution** > **Genetic drift**

**Genetic drift** is a term in population genetics that describes stochastic effects on allele frequencies, mostly due to finite population size.

It is an evolutionary force that, like selection, *reduces* genetic variation. However, drift in one sense can be thought of as *countering* selection, in that the individual's relative fitness (or that of a haplotype, etc) is greatly lessened in those populations where drift operates. Another way of saying this is that if drift is important, the effects of selection are much lessened.

Genetic drift is often exemplified by the population bottleneck, in which a population spends some period of time at low numbers (more precisely, at low effective population size). In small populations, the probability that the most-fit individual will leave the most offspring is greatly reduced (he or she may be struck by lightning, after all). Chance events play a much greater role than they do in large populations. When a small pioneering group form a new population in a new location they bring with them just a small, unrepresentative sample of the genes of their parent population, a phenomenon known as the founder effect.

The stochastic effects of drift can be quantified by considering the probability distribution of an allele's frequency in the next generation. One simple model (the Wright-Fisher model) that does this assumes an hermaphroditic population that sheds gametes into a common pool where random gametes fuse. The question "what is the probability that there will be *i* copies of allele A in the next generation, given that the current allele frequency is *p*?" can be restated mathematically as:
P(*i*) = Pr(*i*|*p*).

Under the Wright-Fisher model, this reduces to the binomial distribution:

P(*i*)=choose(2N, *i*) **p*^*i* * *q*^(2N-*i*)

where choose(2N, *i*) = (2N)!/[*i*! (2N - *i*)!]

From this can be derived the following results:

- the expected frequency of an allele in the next generation is the same as the current frequency: drift does not change the expected allele frequency
- the variance in the frequency of an allele in the next generation, among a large number of replicate populations, is inversely proportional to population size: drift affects allele frequencies by "broadening" their distributions.

The long-term effects of drift, in the absence of other forces, are rather startling. First, random genetic drift will over time cause all populations to become fixed for all alleles. Second, the probability that a given allele fixes as a result of drift is *exactly equal* to its initial frequency in the population.

## ToDo

- historical treatment
- relationship to Kimura's Neutral Theory of Evolution]

## References

- Hartl and Clark,
*Principles of Population Genetics*, chapter 2.

This page is part of the EvoWiki encyclopedia of genetics and molecular biology. Topics: Genetics - Transmission genetics - Molecular genetics - Population genetics - Quantitative genetics - Molecular biology - Genomics |