Table 5.5 A comparison of hypothetical estimates of population subdivision assuming the infinite alleles model using FSTor assuming the stepwise mutation model using RST. Allelic data expressed as the number of repeats at a hypothetical microsatellite locus are given for two subpopulations in each of two cases. In the case on the left, the majority of alleles in both populations are very similar in state. Under the stepwise mutation model the two alleles are separated by a single change that could be due to mutation. The estimate of RST is therefore less than the estimate of FST. In the case on the right, the two populations have alleles that are very different in state and more than a single mutational change apart under the stepwise mutation model. In contrast, all alleles are a single mutational event apart in the infinite alleles model. The higher estimate of RST reflects greater weight given to larger differences in allelic state.

Case 1 Case 2

Subpopulation 1

(number of repeats) 9, 10, 10, 10, 10, 10, 10, 10, 10, 10 9, 10, 10, 10, 10, 10, 10, 10, 10, 10 Subpopulation 2

(number of repeats) 12, 11, 11, 11, 11, 11, 11, 11, 11, 11 19, 20, 20, 20, 20, 20, 20, 20, 20, 20 Allele size variance in subpopulation 1, S1 0.10 0.10

Allele size variance in subpopulation 2, S2 0.10 0.10

Allele size variance in total population, ST 0.947 52.821

Rst 0.789 0.996

Expected heterozygosity in subpopulation 1, H1 0.18 0.18 Expected heterozygosity in subpopulation 2, H2 0.18 0.18 Average subpopulation expected heterozygosity, HS 0.18 0.18 Expected heterozygosity in total population, HT 0.59 0.59

Fst 0.695 0.695

Interact box 5.3 Rst and Fst as examples of the consequences of different mutation models

Under the infinite alleles model allelic state is irrelevant in estimating population structure. However, in the stepwise mutation model, allelic states are weighted in the total estimate of population structure. Computing RST and FST for two subpopulations in a Microsoft Excel spreadsheet will help you develop a better understanding of how mutational models influence the perception of population structure. Enter your own allelic state values in the Excel spreadsheet to explore how allelic state differences as well as allele frequencies produce different estimates of the amount of population structure.

Mutation models for DNA sequences

There are two widely used conceptual models of the process of mutation operating on DNA sequences (note that these types of models also apply in principle to amino acid sequences). One approximation for the process of mutation with DNA sequences is the infinite sites model. Each allele is an infinite DNA sequence and each mutation occurs at a different position along the DNA sequence. The infinite sites model can be thought of as an infinite alleles model built specifically for DNA sequences. A key distinction is that the infinite sites model permits the process of mutation to act on each allele in a population any number of times. As a consequence, the evolutionary "distance" between pairs of alleles can vary since a few or many sites differ between pairs of alleles depending on how many mutations have occurred for each allele. Figure 5.8a shows an example of how mutations might occur for DNA sequences under the infinite sites model. For example, after the fourth base-pair position (or site)

(a) Sequence 1 Sequence 2

0 0

Post a comment