186 CHAPTER 7The dN/dS ratio is a conservative test for adaptation: it often has a value less than 1
even when positive selection has occurred. That is because a single nonsynonymous
change caused by positive selection will be missed if many other nonsynonymous
mutations were deleterious and removed by purifying selection. The next test we
discuss is more sensitive to positive selection, but it requires more data than the
dN/dS ratio.The MK test
A second strategy to find evidence of positive selection builds on the basic idea of
the dN/dS ratio by exploiting data on polymorphism within one of the species. The
MK test is named after John McDonald and Martin Kreitman, who first developed
the approach [25].
Looking back at Equations 7.1 and 7.2, we see that the ratio of the number of dif-
ferences between two species (d) and the heterozygosity within one of the species
(π) will be constant for all DNA sites in the genome that are evolving neutrally. If a
site is evolving adaptively, however, the d/π ratio will be inflated because additional
differences between species will have accumulated by positive selection (making d
larger). Conversely, if a site is under purifying selection, the ratio will be depressed
because selection will prevent many deleterious mutations from becoming fixed
(making d smaller). The MK test assumes that synonymous changes evolve neu-
trally. It compares the d/π ratio at synonymous DNA sites with the d/π ratio at
nonsynonymous sites. If the ratio for nonsynonymous sites is higher, there is evi-
dence of positive selection. If the ratio is lower, there is evidence of purifying (or
balancing) selection. An extension of the MK test was used to make the estimates
cited earlier for the fraction of amino acid differences among species that result
from adaptation and the fraction that result from drift [8, 16].
The MK test has been used to map genes in humans that show evidence of posi-
tive and purifying selection (FIGURE 7.22). In a comparison of 3377 loci in humans
and chimpanzees, one study found that 9 percent of the genes showed evidence
of recent positive selection [6]. Many of these loci fall into interesting functional
groups. Adaptive evolution has been particularly common in transcription factors
and hormone receptor genes. Some of these genetic changes may have been cru-
cial to the emergence of the unique traits we see in modern Homo sapiens.Divergence among populations
In some cases we suspect there has been adaptation to a particular environment,
but don’t know which gene or genes were involved. In this situation we can use a
genome scan to look across the genome for signs of positive selection. One type of
scan compares allele frequencies in two or more populations. Sites in the genome
that have unusually high divergence are suggestive of local adaptation resulting
from the spread of different alleles in different populations.
The Tibetan Plateau is one of the harshest environments inhabited by
humans. It is very cold and very high: many Tibetans live above 4000 m ele-
vation, where oxygen pressure is less than two-thirds what it is at sea level.Futuyma Kirkpatrick Evolution, 4e
Sinauer Associates
Troutt Visual Services
Evolution4e_07.22.ai Date 11-14-2016 01-18-171 2 3 4 FIGURE 7.22 Map of loci on human chromosomes 1–4 where one study found
statistically significant signals of recent selection using the MK test. Each vertical bar
represents a chromosome. In the entire human genome, 304 genes showed signifi-
cant signs of recent positive selection (loci marked in blue), while 813 genes showed
a statistically significant signal of purifying or balancing selection (loci marked in red).
(From [6].)07_EVOL4E_CH07.indd 186 3/23/17 9:09 AM