New Scientist - USA (2021-02-20)

(Antfer) #1

18 | New Scientist | 20 February 2021


SIGNS of a global catastrophe
about 680 million years ago,
known as Snowball Earth, have
been found in the DNA of living
bacteria in the oceans. Their
genomes show that they nearly
died out around this time, says
Haiwei Luo at The Chinese
University of Hong Kong.
Today, tiny photosynthetic
bacteria called Prochlorococcus^
are incredibly abundant in the
surface waters of oceans. A litre
of seawater can contain more
than 100 million of these
cyanobacteria. When Luo and
his colleagues studied their
genomes, the researchers found
that at some point in the distant
past, the most common types
of Prochlorococcus had acquired
many harmful mutations and
lost hundreds of genes.
This shows they went
through what is called a
population bottleneck. When
a population shrinks to a low
number, natural selection is
much weaker and damaging
mutations can accumulate.
The researchers published
these findings in 2017, but were
left puzzling over what caused

this bottleneck. The ancestors
of Prochlorococcus evolved
around 2 billion years ago,
and these bacteria have long
been abundant and widespread.
Only a global catastrophe
could explain it.
Luo and his colleagues
have now worked out that
this bottleneck occurred about
680 million years ago. They did
this with the help of a molecular
clock, which is based on the idea

that, on average, genomes
mutate at a constant rate. The
team estimated that rate based
partly on the ages of fossils
whose appearance suggests
they are ancestors of bacteria
like Prochlorococcus.
That means the bottleneck
occurred during a period of
super ice ages when the planet
got so cold that even the seas
around the equator mostly froze
over, hence the term Snowball
Earth (bioRxiv, doi.org/fvcd).

This would have been a disaster
for Prochlorococcus. “This
explains very well the genetic
evidence,” says Luo.
Some other cyanobacteria
thrive even in polar waters, but
modern Prochlorococcus prefer
the tropics and usually don’t
grow when the temperature is
below 10°C, he says. Yet during
Snowball Earth, a few might
have managed to adapt to the
cold and cling on in refuges
like the briny water in sea ice.
Luo thinks some of the
genetic changes that occurred
at this time are related to cold
adaptation. For instance,
proteins in the cell membrane
that transport substances such
as nitrogen compounds work
poorly in the cold, and several
genes for such proteins were lost.
Instead, the bacteria may
have got the nitrogen they
needed from ammonia, which
can diffuse into cells without
a transporter. The genes
necessary for this were retained.
“I think these interpretations
are reasonable,” says Gregory
Fournier at the Massachusetts
Institute of Technology, who
studies how genomes change
over geological timescales. But
calibrating the molecular clock
with fossil evidence involves
a lot of assumptions, he says.
Paul Hoffman at Harvard
University, who discovered
key evidence supporting the
Snowball Earth hypothesis, also
points to the uncertainties in
the dating. But Snowball Earth
would have left a mark in the
genomes of all the organisms
that survived it, he says. “All
living taxa descended from
Snowball survivors.” ❚

MICROSOFT has developed an
artificial intelligence for its Teams
videoconferencing software that
aims to put people presenting
a remote talk more at ease by
highlighting the most positive
audience reactions.
The AI, named AffectiveSpotlight,
identifies participants’ faces and
uses a neural network to classify
their expressions into emotions
such as sadness, happiness and
surprise, and to spot movements
like head shaking and nodding.
It also uses an eyebrow detection
system to spot confusion, in
the form of a furrowed brow.
Each expression is rated between
0 and 1, with positive responses
scoring higher. Every 15 seconds,
the AI shows the presenter the
person with the highest score
over that time period.
A spokesperson for Microsoft
Research told New Scientist that
“spotlighting audience responses
makes the presenter more aware
of their audience and achieves a
communicative feedback loop”.
The research team declined a
request for an interview.
In a survey of 175 people
conducted by the team, 83 per cent
of those who give presentations
said they often miss relevant
audience feedback when presenting
online, particularly non-verbal
social cues.
To see whether AffectiveSpotlight
could help address this problem,
the team tested it against software
that showed audience members
at random. The AI only highlighted
40 per cent of participants during
talks, compared with 87 per cent
by the random software.
Speakers reported feeling more
positive about doing presentations
with AffectiveSpotlight, although
audience members couldn’t discern
a difference in the quality of
presentation from those using the
AI (arxiv.org/abs/2101.12284). ❚

During “Snowball Earth”,
the planet was mostly
covered in ice

Artificial intelligence Genetics

Chris Stokel-Walker Michael Le Page

RIC

HA

RD

BIZ

LE
Y/S

CIE

NC

E^ P

HO

TO

LIB

RA

RY

News


680m
Years since “Snowball Earth”
almost killed off ocean bacteria

Ancient icy calamity left


its marks on bacterial DNA


AI can tell if people


are enjoying your


video call

Free download pdf