Figure 1. Genetic differences between two homologous chromosome regions. Three single-nucleotide differences are shown: an A/G difference (transition), a C/T difference (transition) and a G/T difference (transversion). Each of these three differences may appear at a different frequency in the population.

Figure 2. Probability that an allele will change in frequency to 100%, given its current frequency. When the selection pressure is small (|s|<~0.01% in an effective population of N_e=10 000), the allele behaves as if neutral (diagonal line). When negative selection is strong (s<–0.1%), it becomes very improbable that the allele will increase in frequency enough to result in polymorphism. Each curve represents a different coefficient of selection, s, conferred by the allele. Positive values of s indicate positive selection, negative values represent negative selection. Probability is calculated according to Kimura (1983) as y=[1–exp(–Sx)]/[1–exp(–S)], where S=4N_es. Therefore we are assuming the effective population size as 10 000, the estimated value during most of human evolution.

Figure 3. A toy example illustrating the evolution of SNPs, seen from (a) the genealogical view and (b) the phylogenetic view. The starting population is genetically homogeneous and the population size is constant. (a) Individuals are shown as large open circles, with two chromosomes (grey bars). Mutations (copying errors) are shown as closed circles on the chromosomes. When variants arising from mutation are faithfully copied into following generations, they are shown as open circles. When a variant reaches 1% in the population, it is shown as a star. In this example, allele a (arising from an erroneous substitution of the parental allele A) is created by a mutation when an individual from the first generation copies its chromosomes and passes allele a to its children. After N generations, allele a has drifted to 1% frequency. In this example, allele b has already arisen as a mutation of allele B on a chromosome with allele a, sometime during the N generations. After an additional M generations, b reaches 1% frequency. The frequency of allele a must be greater than the frequency of allele b, since allele a can be found with both alleles B and b, but allele b is only found with allele a (because it was created on a chromosome with allele a, and is so close to a that it has not been separated by recombination). (b) Each unique DNA sequence is represented by an oval, with branches in the SNP phylogeny represented by arrows. Before N generations have passed, the AB allele is the only unique high-frequency variant. After N generations, allele a reaches 1% frequency, represented as the aB allele (with the mutation labelled Aa on the branch). After M additional generations, the ab allele reaches 1% frequency as a phylogenetic ‘child’ of the aB allele. To help illustrate how phylogenies are represented, we also show an alternate tree that does not correspond to the genealogy shown in (a). This is the tree that would have resulted if allele b had arisen by mutation on an AB chromosome, instead of an aB chromosome.