T HE EvoluTion of GEnEs And GEnomEs 355
molecular evolution. We now understand that the neutral theory is a powerful
explanation for some patterns, such as why pseudogenes evolve so quickly. But
Kimura’s claim that positive selection makes only a trivial contribution to the dif-
ferences between species turns out to have been somewhat exaggerated, as you will
now see.Evolution of coding regions by positive selection
Random genetic drift is responsible for much of the evolution of DNA sequences—
much, but certainly not all. Natural selection occasionally favors changes to a protein,
for example when a species encounters a new environment. Then the typical pat-
tern can be reversed, and DNA mutations that change a protein spread more often
than those that do not. In the 4 million years following the origin of the RNASE1B
gene in the douc langur, nine nonsynonymous mutations became fixed, while only
three synonymous mutations became fixed (see Figure 14.6). Several lines of evi-
dence show that most or all of the changes to the protein produced by RNASE1B
were driven by positive selection that improves its new role in digestion [67].
We can get a rough idea of the relative importance of purifying selection, drift,
and positive selection using a simple statistic based on a comparison of the DNA
sequences of the same gene from two species. We determine dN, the fraction of
sites that differ at nonsynonymous sites (those that do change an amino acid), and
dS, the fraction of sites that differ at synonymous sites (those that do not change
an amino acid). Dividing the first fraction by the second gives us the dN/dS ratio.
Imagine that synonymous and nonsynonymous mutations both have very little
effect on fitness. Then both types of mutations will have the same chance of drift-
ing through the population to fixation, and so we expect dN/dS to be 1. However,
if most nonsynonymous mutations are deleterious and are removed by purifying
selection, then dN/dS will be much less than 1. This is what we see at the great
majority of genes (see F i g u r e 7. 21).
Occasionally loci show a dN/dS ratio greater than 1. This means that more non-
synonymous mutations, which are likely to affect fitness, have been fixed than
synonymous mutations, which are likely to be selectively neutral. That suggests
the nonsynonymous mutations that fixed had a boost from positive selection. The
RNASE1B gene in the douc langur discussed earlier shows exactly this pattern.
So some of the genetic differences we see among species were fixed by adap-
tation, and others by random genetic drift. This raises a fundamental question:
How much do these two processes contribute to evolutionary change? The answer
depends strongly on the group of organisms [11, 25]. About half of the differences
in protein sequences among species of fruit flies (Drosophila), mice (Mus), and
enteric bacteria (E. coli and Salmonella enterica) evolved by positive selection and
half by drift. In our own species, the picture is very different: less than 15 percent
of protein evolution in our recent past has been adaptive.^1 Many genes have been
found in humans that show evidence of positive selection, but they represent only
a small fraction of all the changes that have evolved in the last few million years.
This striking difference between the evolutionary patterns of flies and humans is
a consequence of population size. Drosophila melanogaster has an effective popula-
tion size in the millions, while humans have had an effective population size of only
about 10,000 over much of our evolutionary history. Consequently, genetic drift has
been weaker in D. melanogaster and stronger in humans (see Chapter 7). A deleteri-
ous mutation in the fly that decreases fitness by s = 10–5 will be weeded out by
purifying selection. This prevents it from becoming fixed and contributing to dif-
ferences between species. In humans, however, a mutation with that same selection(^1) The relative contributions of selection and drift to the evolution of differences among species
are estimated using an extension of the MK test that we discussed in Chapter 7.
14_EVOL4E_CH14.indd 355 3/22/17 2:44 PM