62 Science and technology The EconomistJuly 22nd 2017
12 Dr Messer wanted to understand this
process. To do so, he and his colleagues ob-
served the effects of introducing a CRISPR-
Cas9 gene drive they had developed into
fruitflies—insects commonly used in genet-
ic studies.
Their drive’s payload was a gene that
encodes red fluorescent protein, a sub-
stance normally found in a species of sea
anemone. By further genetic tweaking, the
researchers arranged for this protein to be
expressed, in particular, in the insects’
eyes. They thus knew that flies with fluo-
rescent eyes carried their gene drive. Also,
the guide RNAthey selected meant the
drive inserted this payload into the middle
of a gene named “yellow”, thus disrupting
that gene’s action. Flies which inherit a de-
fective version of “yellow” have yellow
bodies, rather than black ones.
After experiments with thousands of
flies, Dr Messer and his team found that the
gene drive successfully inserted itself into
the insects’ DNAabout half the time, pro-
ducing flies with fluorescent eyes and yel-
low bodies. The other half of the insects
nearly all had yellow bodies, but did not
have fluorescent eyes. That indicated
CRISPR-Cas9 had cut the DNAin the right
place, thus disrupting the function of “yel-
low”, but had failed to insert itself into the
incision. Sequencing the genomes of these
flies confirmed that in virtually all cases
the consequence was a mutation that ren-
dered the flies (and would have rendered
any offspring) resistant to the gene drive.Red signal
The team then repeated their experiments
with fruitfly lines from five continents.
They found that the proportion of flies be-
coming resistant to the drive varied from
about 60% down to 4%. Differences in resis-
tance to the drive were not caused by any
initial differences in the target sequence.
That did not vary between the five lines.
They must therefore have stemmed from
other (as yet unknown) genetic differences
between the flies. This is a worry because it
suggests that even if a drive works well in
laboratory animals, it may fail in the wild
when it encounters populations with high-
er resistance. Getting to the bottom of what
is causingsome lines to be more resistant
than others will be an important step to-
wards the development of gene drives that
can spread traits through a species, Dr
Messer reckons.
Two further modifications ofCRISPR-
Cas9 gene drives may help. The first is to
equip them with several guide RNAs, al-
lowing Cas9 to cut chromosomes at more
than one place. An organism would have
to develop resistantDNAsequences at all
of these to become fully immune to the
drive. Dr Messer and his colleagues have
made such a drive, containing two guide
RNAs, and have found that it did indeed
lower the proportion of flies that devel-oped resistance—though estimates made
with computermodelssuggest this is still
not enough for the drive to reach more
than about half the population.
The second approach is to put the drive
into the middle of a gene that, unlike “yel-
low”, an organism needs to survive. This
might be expected to disrupt the gene and
kill the organism. But if the inserted DNA
has, at one end, a replica of the part of the
disrupted gene that hasbeen displaced by
the insertion, this can meld seamlessly
with its counterpart in the animal, preserv-
ing the gene’s function and Cas9’s abilityto recognise it. If the join is not seamless,
though, the gene will fail and the animal
will die. Mutant genesresistant to the drive
will thus be unable to spread.
At least two groups, one based at Impe-
rial College, in London, and the other at
Harvard University and the Massachusetts
Institute of Technology, are working on
drives aimed at essential genes in mosqui-
toes and which use multiple guide RNAs. If
they succeed, they will breathe new life
into the field. As Dr Messer observes, “it is
difficult to cheat evolution.” Whether it is
impossible to do so remains to be seen. 7B
EAUTY, proverbially, is in the eye of the
beholder. But surroundings matter. A
paper published two years ago in Nature
found a correlation between people’s
sense of well-being and the “scenicness”
of where they lived. The paper’s authors
measured scenicness by asking volunteers
to play an online game called Scenic-or-
Not, which invites participants to look at
photographs of neighbourhoods and rate
their scenic value on a scale of one to ten.
The correlation, the paper’s authors
found, held true whether a neighbour-
hood was urban, suburban or rural. It bore
no relation to respondents’ social and eco-
nomic status. Nor did levels of air pollution
have any influence on it. The authors also
discovered that differences in respondents’
self-reported health were better explained
by the scenicness of where those respon-
dents lived than by the amount of green
space around them.
Pinning down what scenicness actuallyis, though, has always been a frustrating ex-
ercise for scientific types. The team behind
thatNaturepaper, Chanuki Seresinhe and
her colleagues at Warwick Business
School, have nevertheless decided to have
a go. And they think they have succeeded.
As they report in Royal Society Open Sci-
ence, they have adapted a computer pro-
gram called Places to recognise beautiful
landscapes, whether natural or artificial,
using the criteria that a human beholder
would employ.
Places is a convolutional neural net-
work (CNN), a type of program that can
learn to recognise features in sets of data,
such as images, presented to it. CNNsoften
form the basis of face-recognition soft-
ware. Places, though, as its name suggests,
is optimised to recognise geographical fea-
tures. Ms Seresinhe and her team taught
the program to identify such things as
mountains, beaches and fields, and va-
rious sorts of buildings, in pictures present-Artificial intelligenceAdmiring the scenery
Computer analysis of what people find scenic may help town plannersCan you tell a green field from a cold steel rail?