This graph shows the results of natural selection and random drift in a diploid population. The thick purple line shows the predicted change in frequency of the red allele due to natural selection alone. It represents what would happen in an infinitely large population. The thin red lines are simulations that include both random drift and natural selection.
Set the relative fitness of one genotype to 1.0, and set the relative fitnesses of the other genotypes to less than or equal to 1.0. When all three fitnesses are 1.0, only random drift will operate; the results should be the same as on the previous web pages that showed the effects of random drift.
Selection plus drift: diploid
First, try making the red allele homozygote (rr genotype) have a fitness of 1, with the other two genotypes having fitnesses equal to each other and less than 1. You should see that the red allele increases in frequency--there is directional selection favoring the red allele. If the selection coefficients against the rb and bb genotypes are small--they have fitnesses close to 1--random drift becomes more important. Try different population sizes, to see that random drift also becomes more important as the population size gets smaller.
If the rr genotype has a fitness of 1 and the rb and bb fitnesses are equal to each other and less than 1, that means that the b allele is dominant. Note that dominant does not mean better, or more common; it just means that the rb genotype is like the rr genotype in fitness. Keeping the population size and the fitnesses of the rr and bb genotypes constant, try different fitnesses for the rb genotype. Try it with fitness equal to the bb genotype's fitness (b is dominant), with fitness equal to the rr genotype (r is dominant), and with fitness halfway in between that of the rr and bb genotypes (the alleles are codominant). You should see that the dominance has a large effect on the rate of increase in frequency of the favored allele. Try a few different initial allele frequencies; the effect of dominance is especially important when the favored allele is rare and recessive.
Next, try making the heterozygote (rb genotype) have a fitness of 1, with the other two genotypes having fitnesses less than 1. This is balancing selection. You should see the allele frequency reach an equilibrium. Try different fitnesses for the two homozygotes and see how that changes the equilibrium allele frequency.
Finally, try selection coefficients that are around 1/2N. For a population size of 50 diploid individuals, this would be relative fitnesses of 1 and 0.99. You should see that random drift becomes the most important factor, so that the results are about the same as with the pure random drift case (when all three fitnesses are 1).
This page was last revised March 6, 2009. Its address is http://udel.edu/~mcdonald/evolseldriftdip.html.
©2009 by John H. McDonald. You can probably do what you want with this content; see the permissions page for details.