## Interact box Genetic drift

Genetic drift can be simulated with either Populus or PopGene.S2. Try using Populus and following the instructions here.

Under the Model menu, select Mendelian Genetics and then Genetic Drift. Make sure the Monte Carlo tab is selected. The simulation dialog has entry fields for Population Size (N) and Number of Loci (or replicates). Number of generations shown in the graphs can be specified as 3N or a fixed number of generations (Other:) with radio buttons (the program will automatically set the number of generations if all loci go to fixation or loss before the fixed number is reached). Set allele frequencies of all loci collectively. Press the View button to see the results graph. The model can be rerun with the Iterate button in the results window.

Set the parameters for 10 loci and an initial allele frequency of 0.5, and view the results for 800 generations so that all fixation/loss events are visible. Examine drift for population sizes of 4, 20, 50, and 100. Record the generation of fixation or loss for 40 replicates of each population size. What is the average number of generations to fixation or loss? Are there equal numbers of fixation and loss events?

Progressively reduce (or increase) the initial allele frequencies by intervals of 0.1 for a single population size. Record the generation of fixation or loss for 50 replicates of initial allele frequency. What is the observed relationship between initial allele frequency and probability of fixation or loss? Do these averages change if the population size changes?

Hint: using a spreadsheet program like Microsoft Excel can speed the calculation of averages for a list of values. Enter the values in columns and then use the average function ("=AVERAGE( )" with the range specified in the parentheses, such as "C1 : C10" to indicate the values of cells 1 through 10 in column C). A wide range of other useful functions is provided under Functions ... in the Insert menu.

The next section of the chapter will develop two probability models of sampling error to provide more rigorous evidence for the conclusions reached in this section and to reach additional generalizations about the action of genetic drift.