SCIENCE sciencemag.org 4 JANUARY 2 019 • VOL 363 ISSUE 6422 23
perhaps, Przeworski speculates, because
the variants curb older men’s fecundity.
Or perhaps the hypothesized benefit that
healthy grandmothers confer on grandchil-
dren was at work.
The researchers did find two genes that
suddenly became rare at older ages, suggest-
ing they were harmful. One was ApoE 4 : As
expected, fewer carriers—especially women—
lived past age 80. Also, fewer men with a vari-
ant of the CHRN 3 gene that makes it harder
to quit smoking survived past the age of 75
than did men without the variant.
The researchers concluded that natural
selection has not yet had time to eliminate
these two alleles, perhaps because changes
in the environment and human behavior
only recently made them deadly (Science,
20 May 2016 , p. 876 ). For example, the
CHRN 3 allele wouldn’t have affected sur-
vival until many men were smoking. And
women who were more active in the past
might have been less vulnerable to the
cardiovascular diseases caused by
ApoE 4 , Przeworski speculates.
The researchers spotted another
intriguing pattern. Genetic variants
that lead to early puberty also be-
came rarer in older age groups. Nat-
ural selection may have preserved
those variants even though they shorten life
span because they also boosted fertility.
A LONG LIFE, though, is much less important
to evolution than fertility. When it comes
to the game of evolution, in fact, the per-
son who has the most kids wins by pass-
ing on the most genes. With the advent of
birth control, people in industrial societies
have more control than ever over their own
fertility—but new studies zeroing in on the
genes underlying fertility show the forces of
selection may still be at work.
Multiple studies have suggested that when
food sources became more reliable in indus-
trialized societies, women began to mature
faster, weigh more, give birth to their first
child earlier, and enter menopause later—all
traits possibly linked to having more babies.
But researchers have been unable to tie those
trends to underlying genes to get direct evi-
dence of natural selection. Quantitative ge-
neticist Peter Visscher and his colleagues at
the University of Queensland in Brisbane,
Australia, realized they could use the UKB
to see firsthand which gene variants under-
lie those traits in people today, and whether
they are really linked to fertility.
They searched the UKB’s full cohort for
people who had the most babies to see what
traits they share, and what genes correlate
with those traits. They documented the
number of live births for women over age
45 and men over age 55. Then, they analyzed
traits in women and men that might have in-
fluenced fertility, such as age of first birth,
age of menopause, height, weight, body
mass, blood pressure, and education. They
found 23 traits in women and 21 in men
linked to having more children. Not surpris-
ingly, mothers who gave birth early and had
late menopause—and therefore had a longer
reproductive span—were more fertile. So
were women who were heavier and shorter,
perhaps because shorter bodies are more
energy efficient, leaving a bigger reserve for
pregnancy and nursing.
Visscher and his colleagues then set out
to identify the genetic basis of these fertility-
linked traits. They analyzed data from
157 , 807 of the women and 115 , 902 of the
men. As predicted, they found that the most
fertile women had higher frequencies of al-
leles that tend to make them shorter and
heavier. In men, greater fertility was asso-
ciated with more alleles that contribute to
a higher body mass index and hand-gripstrength. That suggests men with genes that
make them taller and bulkier have more kids
than sedentary types, whether because of fe-
male choice, some health-related reason, or
the men’s own preference.
Not all traits linked to fertility are physi-
cal or likely to have a big genetic component:
Among women who had their first child later
in life, those who had more education and
did better on an intelligence test had more
babies. This may be because better-educated
couples tend to be wealthier and can afford
more children.
But the fact that genes linked to traits
thought to increase fertility are indeed more
common in fertile people backs up the idea
of recent selection on our genomes, even as
both the environment and humans’ prefer-
ences for mates and families are changing.
“The UK Biobank allows us to show that nat-
ural selection not only took place in the past,
but it’s still ongoing,” Visscher says.TEASING OUT natural selection from other
factors shaping genes can be tricky, how-
ever, especially when multiple genes work
together to influence complex traits, such
as height. About 5000 gene variants si-
multaneously influence a person’s height,
some boosting it, some reducing it, says
Jian Yang, a statistical geneticist at the
University of Queensland. The UKB’s huge
database allows researchers to find new
variants and explore their impact and origins.Using other databases, researchers had
found that the number of genes that con-
tribute to tallness in Europeans increased
on a cline from south to north. Many re-
searchers, including Berg, had concluded
that northern Europeans had inherited
those genes from an ancient migration—
that of the Yamnaya herders who migrated
from the Eurasian steppe to central Europe
about 40 00 years ago. Berg and others sug-
gested natural selection had favored tall-
ness in the Yamnaya or their ancestors,
and ancient DNA reveals that the Yamnaya
were tall.
But now, with UKB data, population ge-
neticist Graham Coop of UC Davis and his
colleagues, including Berg, are challenging
that finding. In a bioRxiv preprint posted in
June 2018 , they analyzed genetic and height
data on 50 0, 000 people from the 2017 UKB
data release. With so many people from
similar backgrounds, the researchers could
identify more height alleles, as well as note
differences in diet, disease, and
the environment. They found that
northerners had no more tall vari-
ants than southerners.
“It’s true people in northern
Europe are taller on average, but
there is no evidence this has any-
thing to do with natural selection,” Berg
says. He speculates that northerners’ height
might be an environmental effect, perhaps
from a diet richer in protein, or from fewer
childhood or prenatal illnesses.
Although UKB data cast doubt on natural
selection’s role in that case, they do suggest
that evolution has favored genes for short-
ness in pygmy populations on the island of
Flores in Indonesia. Visscher and colleagues
scanned the DNA of Flores people for genes
the UKB had linked to short stature. They
found that Flores pygmies carry more such
gene variants than their closest relatives in
New Guinea and East Asia, suggesting evo-
lution favored genes for shortness on the
island (Science, 3 August 2 018 , p. 439 ). All
these studies have generated “huge buzz
among evolutionary biologists about how
biobanks can provide very deep informa-
tion about the genetics of different popula-
tions and their evolution,” Kelso says.
She hopes to work with researchers de-
signing databases in Africa and Asia to
identify archaic DNA in those populations.
Thanks to the success of the Neanderthal
work, many researchers are eager for data
from Melanesians, because they have inher-
ited traces of DNA from Denisovans—the
mysterious cousins of Neanderthals who
lived in Siberia more than 50,000 years ago.
“That would be amazing, to get Denisovan
DNA from more living people [in biobanks].
That’s our dream,” Kelso says. j“No one was thinking about Neanderthal
traits when we designed the protocol.”
Rory Collins, University of OxfordPublished by AAASon January 3, 2019^http://science.sciencemag.org/Downloaded from