EVoluTioN iN SPACE 203
genetically very similar. In short, for a selectively neutral locus, a single migrant
per generation is sufficient to prevent drift from causing two populations to diverge
very much. Remarkably, this conclusion is independent of the population size. (Of
course, the picture is very different if selection is at work, as we saw earlier.)
This theoretical fact suggests a way to estimate the amount of migration between
populations. The idea is to use neutrally evolving genes to estimate FST. As explained
in Chapter 7, neutral genes can also be used to estimate Ne. (The heterozygosity, π,
is proportional to Ne and the mutation rate μ. By estimating π from genetic data, and
knowing μ, E q u a t i o n 7.1 can be used to estimate Ne.) Given values for FST and Ne,
we find the value of m that solves Equation 8.5. That is the basic strategy that was
used to estimate the migration rate in the pocket mouse study described earlier [12].
This approach is one of several indirect methods to estimate migration using
genetic data. These methods enjoy several advantages over direct methods like those
used in the lizard study shown in Figure 8.5. Indirect methods average gene flow
over many generations, and are sensitive only to migrants that actually contribute
to genetic mixing between the populations. (A limitation of Equation 8.5 is that it is
based on assumptions that are often violated in nature, but more sophisticated meth-
ods have been developed that relax those assumptions.)
Drift also causes populations that live in continuous habitats to diverge, which
can result in a pattern of isolation-by-distance. In many cases, however, distant
populations are more genetically similar than would be expected from the amount
of gene flow that they currently experience. One explanation is that the popula-
tions may not be at an equilibrium. Earlier we saw that humans show a pattern of
isolation-by-distance (see Figure 8.7). FST increases with distance, but populations
living on different sides of the globe are still genetically very similar despite the
absence of gene flow before the twentieth century. That pattern reflects the history
of how humans colonized the planet some 100,000 years ago, rather than an equi-
librium between current gene flow and drift. A second factor that can genetically
homogenize a species over large spatial scales is a history of frequent extinction
and recolonization. In some species, populations that are wiped out by distur-
bances (such as fires) are replaced by colonists that move into the empty habitat.
This generates bouts of high gene flow that decrease divergence, depressing FST
much below what would be expected from typical rates of movement.Gene flow, local adaptation, and drift
We have seen that two different evolutionary processes—selection and drift—
can cause populations to diverge. This can make it challenging to decide whether
genetic differences between populations are signs of local adaptation or simply the
result of neutral drift.
It is now possible to sample many genes or even entire genomes from differ-
ent populations of some species. How can those data be used to hunt for genes
involved in local adaptation? The most basic approach is simply to scan the genome
for regions that show unusually high FST between two populations. The idea here
is that neutrally evolving regions of the genome will show a baseline level of FST
caused by drift, and regions that show much higher divergence may be under local
adaptation. Further evidence for local adaptation at these candidate regions can
be gleaned using comparisons with additional populations, and by finding genes
within the regions that are plausible targets of selection. In the three-spined stick-
leback that we discussed earlier, comparisons between three independent pairs
of stream and marine populations show repeated peaks of high divergence (see
Figure 8.8). One of the highest peaks is on chromosome 4 and corresponds to the
location of the Eda locus that controls the striking differences between the popula-
tions in lateral bony plates. This is compelling evidence for local adaptation at both08_EVOL4E_CH08.indd 203 3/23/17 9:12 AM