Skip to main content
Helicase

Evolution Lab

Living things vary, and some of that variation is passed on. Over generations, two things change how common each version of a gene is: natural selection, where versions that help survival and reproduction spread, and genetic drift, which is pure chance.Allele frequencies change from one generation to the next. Natural selection makes alleles that improve survival and reproduction more common, while genetic drift changes frequencies at random, because only some individuals happen to pass on their genes. Drift is strongest in small populations.This lab runs the Wright-Fisher model: each generation, selection shifts the expected frequency of allele A according to the fitness of each genotype, and drawing the next generation's gene copies at random adds drift. Compare replicate populations with the expectation without drift, and with Kimura's probability of fixation.Source: NHGRISource: Wright, 1931Source: Kimura, 1962

EvolutionGeneticsEcology

Simplified model, not a forecast

One gene with two alleles, A and a, in a population of fixed size that mates at random, with one generation replacing the last. Mutation and migration are left out. Real populations are messier, but the same two forces, selection and drift, are at work in all of them.Source: Wright, 1931

Start from a scenario

Frequency of allele A over the generations, in each population and on average
  • Average of the populations
  • Without drift
After 100 generations, allele A is fixed in 3, lost in 4 and still varying in 1 of 8 populations. Average frequency 49%; without drift 50%. Use the left and right arrow keys to read values.0%20%40%60%80%100%0102030405060708090100GenerationFrequency of allele A

After 100 generations

Fixed (only A left)
3
Lost (only a left)
4
Still varying
1

Average frequency of A: 49%. Without drift it would be 50%.

Theory: with these settings, A eventually takes over in about 50% of populations (Kimura, 1962).

Drift in a small population

Established

In the model: There is no selection at all, yet allele A wanders up and down in every population, and within a few dozen generations some populations lose it entirely or keep nothing else.

In real populations: Chance alone changes allele frequencies from one generation to the next, and the smaller the population, the faster it does so. This is genetic drift.

Sources

Population
20 individuals
50%
Selection
None
Run
100
8

After 100 generations, allele A is fixed in 3, lost in 4 and still varying in 1 of 8 populations. Average frequency 49%; without drift 50%.

Selection and drift

Natural selection has a direction: alleles that help their carriers survive and reproduce become more common. Darwin’s finches and the peppered moth are well-studied examples.Source: NHGRISource: Grant & Grant, 2006Source: Cook et al., 2012 Genetic drift has no direction: it is the change that comes from chance, because only some individuals happen to pass on their genes. It is strongest in small populations.Source: Wright, 1931

The two interact. Kimura worked out the probability that an allele is eventually fixed, carried by every member of the population, given its advantage and the population size. With no advantage, that probability is simply the allele’s starting frequency; a slight advantage raises it only a little, so most new beneficial alleles are still lost by chance.Source: Kimura, 1962

How each generation is made

  1. Selection: each genotype leaves offspring in proportion to its fitness. AA has fitness 1 plus the selection setting, aa has fitness 1, and Aa sits between them by the chosen dominance (recessive, additive or dominant).
  2. Drift: the next generation’s gene copies, two per individual, are drawn at random from the frequencies after selection. In small populations this sampling shifts frequencies a lot; in large ones, hardly at all.
  3. The line marked “without drift” repeats step 1 alone, as if the population were infinitely large.

The same settings always give the same result until you choose “Run again with new chance”. Large populations use standard approximations to the random draws, which are close enough for this purpose.

Your progress

Sources & further reading

The explanations on this page are our own summaries of these sources. Follow a link to read the original; if anything here disagrees with it, the original wins. How we choose and check sources.