If you have ever heard of the claylike mineral known as zeolite, chances are you share your home with a cat. You may also know that it comes in a powder, and that it is good at wicking out liquids and smells—ideal for concealing the minor indignities of being a feline. Desirée Plata, a civil engineering professor at MIT, uses zeolite for a different kind of molecular cleanup: Combine it with a metal catalyst—in Plata’s case, copper—add some heat, and it will trap and destroy methane, one of the most potent greenhouse gases.

Methane is a quixotic warming agent. Unlike carbon dioxide, which persists in the atmosphere for thousands of years, natural forces remove it within roughly a decade, mostly when it reacts with other molecules in the air. But for the brief time methane mixes aloft, it punches far above its weight, producing 80 times the warming effect of carbon dioxide over 20 years. By some estimates, it has been responsible for a third of anthropogenic warming so far, despite receiving far less attention. It is also notoriously difficult to track where the gas comes from. Some methane is trapped underground and then uncorked by natural fissures or by people boring into the ground for oil—or for methane itself, under the more anodyne name “natural gas.” But it can also be created anew by microbes wherever there’s a lot of biomass and very little oxygen: rice paddies, landfills, wetlands, or inside the digestive tracts of cows.

Over the past few years, the atmospheric concentration of methane has been spiking, puzzling and alarming climate scientists. According to the National Oceanic and Atmospheric Administration, the measurements from 2021 are poised to show the biggest increase since scientists started consistently measuring the gas. (The data takes a few months to catch up.) Is it a blip or a sustained rise caused by certain emissions sources? Or perhaps something else has changed in the cocktail of atmospheric gases, so that methane is destroyed less readily than before? “‘I don’t know’ is the honest answer,” says Rob Jackson, a climate scientist who studies methane at Stanford University. “The concentration increases are frightening. And if they continue, this is terrible news.”

 

What’s clear is that the world’s first priority needs to be cutting methane emissions, Jackson adds. Sometimes that’s as simple as turning a screw on a leaky pipeline valve or plugging up a defunct gas well. But there are limits to that pinpointed strategy. With CO2, zeroing in on a so-called “super emitter” is as simple as scanning the horizon for the smokestacks of a coal-fired power plant. But comparable sources of methane emissions are often more sporadic—a pipeline leak here, a landfill plume there—a game of whack-a-mole for environmental watchdogs inhibited by limited surveillance. Accountability is also tricky: The methane emissions of a particular herd of cows can’t be measured as consistently as the CO2 spewed by a freeway full of cars.

Natural emissions, which are estimated to be about 40 percent of methane emissions, are even trickier, and they are likely to accelerate as the world warms, in part by firing up gas-emitting microbes that live in permafrost or underneath sea ice. “The problem with natural emissions is that there is not a lot we can do with them,” Jackson says. “It’s hard to estimate the emissions of the Chesapeake Bay, or more terrifyingly, measure what will happen if the Arctic starts melting. That’s letting the genie out of the bottle, and it’s impossible to get it back in.”

So perhaps, Jackson and other scientists suggest, it's time to think about removing methane from the atmosphere, in addition to cutting back on new emissions. It’s an idea that’s far more advanced for carbon dioxide—and perhaps for good reason, given that CO2 is the leading cause of warming and that humanity will be living with today's CO2 emissions for thousands of years. But with methane, proponents argue there’s a rationale for swift action—a chance to return to preindustrial levels within decades, thanks to its short life span. Jackson and other scientists have argued that the heating effects of methane are chronically undervalued, because current climate policies emphasize long-term temperature goals that extend far beyond the lifetime of a methane molecule. The value of reducing methane levels spikes when you factor in the benefits of preventing warming now.

But the idea presents a paradox: There may be too much methane coming from everywhere to do that with emission cutbacks alone, Jackson says. Yet perhaps there’s not enough of it in the air to feasibly take it out.

Destroying methane is, in a chemical sense, a relatively easy task. Nature does it constantly. Methane is a single carbon atom surrounded on four sides by hydrogen, and these bonds are broken apart by a process called oxidation—it involves oxygen atoms plus some goading by energy and chemical catalysts.

There are many ways to make that chemical reaction happen, explains Renaud de Richter, a scientific adviser to Methane Action, a nonprofit that advocates removing the gas from the environment. Most methane is oxidized naturally when it reacts with either chlorine atoms or hydroxyl radicals in the atmosphere. So one idea is to spray iron salts, perhaps with the assistance of cargo ships, that will coax more chlorine atoms out of briny ocean air. The idea is being tested by researchers at the University of Copenhagen inside a laboratory gas chamber. Another concept, favored by de Richter, would involve using thermal towers that passively suck in air and break down methane through photocatalysis—a process involving sunlight and metal catalysts. But neither idea has been tested in the real world yet.

Compared to those methods, Plata’s zeolite-based removal is old science. The idea emerged from the industry that produces methanol—a liquid chemical that’s used as an antifreeze, among other things, and can be converted into a variety of fuels like ethanol. Microbes do this naturally. Sitting in the seabed, they get a little methane from the ground below, a little oxygen from the air up above, and produce methanol. Industrial players have tried to replicate this using a powdered zeolite as a sort of “molecular sieve” that traps the methane in its pores, then oxidizes the molecules with heat, oxygen, and metal catalysts. 

For methanol producers, the problem with this setup is that the reaction is difficult to control precisely. The concoction tends to over-oxidize, turning the precious methanol into carbon dioxide and water vapor. “Engineers have obsessed over how to prevent that from happening,” Plata says. Their solution is to try out funky modifications, like alternately flooding the reaction with methane and oxygen, although this makes the whole process inefficient.

But Plata’s goal isn’t to produce any methanol; it’s simply to get rid of methane. Hence her solution: Don’t sweat the carbon dioxide. “People freak out when I say that,” Plata says. Yes, it’s a little odd to suggest turning one greenhouse gas into another. But, she says, because of CO2’s substantially lower warming effect, its relative effect on the climate is “a minuscule little blip” compared to letting methane hang around. In the lab, Plata dried up a copper and zeolite mixture and placed it in a tube with various mixtures of atmospheric gases, including methane at varied concentrations. “It works. I’ll say that. We can convert low levels of methane,” Plata says. “The question is about how quickly you can make it work.”

Last month, Plata’s team received a $2 million grant from the US Department of Energy intended to quickly move the technology out of the lab. The next step is converting their powdered catalyst into a zeolite-based filter that’s easier to push air through—a process she compares to the catalytic converter at the back of a car. Plata wants to install the filters in places where methane is concentrated, but there’s not enough of it to burn—a process known as flaring that’s commonly used to get rid of methane leaking from natural gas and oil wells. Flaring and other thermal techniques can destroy methane at concentrations as low as 2,000 parts per million. She imagines the zeolite-based filtration being used in less concentrated environments, like mine shafts or indoor dairy farms.

Jackson, who is involved in a team working on similar technology, says he likes the zeolite strategy because it occurs “inside a box.” Unlike spraying chemicals into the open air from cargo ships, it’s easier to count how much methane is being destroyed and whether there are knock-on effects—more reactions happening in the air producing byproducts that you might not want. But he acknowledges that “zeolites aren’t magic.” Among the key concerns are producing a material that allows as much air as possible to flow through it, and bringing down the temperature of the reactions—both in an effort to conserve energy. (Though an improvement over other methods, Plata’s lab process works best at a balmy 300 degrees Celsius.)

But Jackson feels the concept is getting close to viability for what he describes as a sweet spot—for deployment to places where there isn’t good technology to destroy or pull back on high volumes of methane emissions. Over time, Jackson says, the idea is to reduce the concentration of methane that can be viably filtered to about 2 parts per million—the background level of the gas in the atmosphere.

Getting there will be tough, says Klaus Lackner, a geophysicist at Arizona State University and a pioneer in CO2 capture technology. Lackner says the challenges of removing either gas from the atmosphere are similar, but in his view, far more daunting to scale for methane. Atmospheric concentrations of methane are relatively low—CO2, by comparison, is at 412 parts per million—and it would be almost impossible to push enough air through enough filters to trap a worthwhile amount, he says. What’s more, even if you could, “you have to pull a lot out to make a difference,” because natural processes may compensate to replace the artificially removed methane on a global scale. The oceans release a little extra, the microbes munch a little less. “You have to be as big as nature,” he says.

The barrier to being as big as nature is the potential cost. It would require either a tremendous number of passive machines to process enough of the atmosphere to make a dent in overall concentrations, or energy-hogging fans that are “barely affordable” even for CO2 capture. (By one estimate, humanity would need to build some 10,000 direct air capture facilities by the end of the century to actually reduce CO2 levels.) But, he adds, it’s useful to keep studying the various strategies. “At the end of the day, you could say we have an insurance policy,” Lackner says—especially if methane emissions start to get truly out of control, such as if a warming Arctic causes a runaway release of the gas.

Jackson acknowledges the challenges—and that reducing methane emissions, as well as CO2, is priority number one. That means tighter rules for emitters and better surveillance to identify where the gas is escaping. Currently, “it’s not clear who would pay” for methane removal, he says, even in cases where the technology would be sited at businesses with high emissions, like at a dairy farm or a coal mine.

But that could change, he adds. The idea of methane removal is getting more attention, including an analysis of various technologies in the 2021 report from the Intergovernmental Panel on Climate Change. He also points to the recent signing of more than 100 nations to the Global Methane Pledge, a US- and EU-led effort to cut emissions by 30 percent from 2020 levels by the year 2030. That may encourage nations to consider more stringent rules for facilities that produce methane, and perhaps put a higher value on the carbon equivalence of methane emissions, translating into incentives for removal or taxes for emissions. The point, Jackson says, is that we should do everything we can to stop warming now: “Methane is the strongest lever we have for that.”

The inside of a tokamak—the doughnut-shaped vessel designed to contain a nuclear fusion reaction—presents a special kind of chaos. Hydrogen atoms are smashed together at unfathomably high temperatures, creating a whirling, roiling plasma that’s hotter than the surface of the sun. Finding smart ways to control and confine that plasma will be key to unlocking the potential of nuclear fusion, which has been mooted as the clean energy source of the future for decades. At this point, the science underlying fusion seems sound, so what remains is an engineering challenge. “We need to be able to heat this matter up and hold it together for long enough for us to take energy out of it,” says Ambrogio Fasoli, director of the Swiss Plasma Center at École Polytechnique Fédérale de Lausanne in Switzerland.

That’s where DeepMind comes in. The artificial intelligence firm, backed by Google parent company Alphabet, has previously turned its hand to video games and protein folding, and has been working on a joint research project with the Swiss Plasma Center to develop an AI for controlling a nuclear fusion reaction.

In stars, which are also powered by fusion, the sheer gravitational mass is enough to pull hydrogen atoms together and overcome their opposing charges. On Earth, scientists instead use powerful magnetic coils to confine the nuclear fusion reaction, nudging it into the desired position and shaping it like a potter manipulating clay on a wheel. The coils have to be carefully controlled to prevent the plasma from touching the sides of the vessel: this can damage the walls and slow down the fusion reaction. (There’s little risk of an explosion as the fusion reaction cannot survive without magnetic confinement).

But every time researchers want to change the configuration of the plasma and try out different shapes that may yield more power or a cleaner plasma, it necessitates a huge amount of engineering and design work. Conventional systems are computer-controlled and based on models and careful simulations, but they are, Fasoli says, “complex and not always necessarily optimized.”

DeepMind has developed an AI that can control the plasma autonomously. A paper published in the journal Nature describes how researchers from the two groups taught a deep reinforcement learning system to control the 19 magnetic coils inside TCV, the variable-configuration tokamak at the Swiss Plasma Center, which is used to carry out research that will inform the design of bigger fusion reactors in the future. “AI, and specifically reinforcement learning, is particularly well suited to the complex problems presented by controlling plasma in a tokamak,” says Martin Riedmiller, control team lead at DeepMind.

The neural network—a type of AI setup designed to mimic the architecture of the human brain—was initially trained in a simulation. It started by observing how changing the settings on each of the 19 coils affected the shape of the plasma inside the vessel. Then it was given different shapes to try to re-create in the plasma. These included a D-shaped cross section close to what will be used inside ITER (formerly the International Thermonuclear Experimental Reactor), the large-scale experimental tokamak under construction in France, and a snowflake configuration that could help dissipate the intense heat of the reaction more evenly around the vessel.

Ten years ago, a company called Agrobot demonstrated a strawberry-harvesting robot in a field in Davis, California. Today, Agrobot’s strawberry picker remains a prototype.

The long wait underscores the challenge for any berry-picking robot: Identify a berry that is ripe enough to pick, grasp it firmly but without damaging the fruit, and pull hard enough to separate it from the plant without harming the plant. Agrobot CEO Juan Bravo said his company’s machine can’t compete with people who can pick fruit by hand and pack it into clamshells.

Still, growers are looking ahead to a day when it will be hard to find people willing to stoop in the fields all day, and expensive to pay them. So growers, technologists, and researchers are continuing to pursue machines that can do the job. A recent survey of nearly 50 robotic harvesting projects showed that strawberry-picking projects attracted more interest than projects targeting any other fruit over the past two decades.

In the latest sign of this interest, indoor-farming company Bowery recently acquired Traptic, a Silicon Valley startup created in 2016 that last year began commercial deployments with Naturipe and Blazer Wilkinson, two large strawberry growers. Bowery will adapt Traptic for indoor vertical farming because its systems, like most of its competitors’, primarily operate outside in the fields of California or Florida.

Digital twins—virtual representations of real-world things—are already a mainstay in manufacturing, industry, and aerospace: There are digital doppelgängers of cities, ports, and power stations. The term was first introduced in 2010 by  NASA researcher John Vickers in a report about the agency’s technology road maps, and industry analysts estimate the market for digital twins could reach nearly $50 billion by the year 2026. 

It wasn’t long before the idea crept into biology. In 2016, Bill Ruh, then-CEO of GE Digital, predicted that “we will have a digital twin at birth, and it will take data off of the sensors everybody is running, and that digital twin will predict things for us about disease and cancer and other things.” A digital twin could inform tailored treatments for a patient and predict how their disease might develop. It could even be used to trial potential treatments, rather than testing them on the patient—a process that can be filled with risk. 

So far, these projects are mostly in their early stages. A research program called Echoes, involving researchers in Europe, the United Kingdom, and the United States, is working to build a digital heart. Siemens Healthineers, a German medical device company, is aiming to do the same. Dassault Systèmes, a French software company, teamed up with the US Food and Drug Administration to approve what it calls “The Living Heart.” Austrian company Golem is creating digital twins of vulnerable people who live alone. The idea is that the digital twin continuously monitors their health, alerting caregivers if they fall ill and need help. 

Now researchers are shooting for the loftiest goal: to twin the brain. Neurotwin, an EU-funded project, wants to design a computerized model of an individual patient’s entire brain. 

The Neurotwin team is hoping the model can be used to predict the effects of stimulation for the treatment of neurological disorders, including epilepsy and Alzheimer’s disease. They’re planning a clinical trial that will kick off next year and create digital twins of about 60 patients with Alzheimer’s, who will receive a brain stimulation treatment that has been optimized specifically for their brain. A second clinical trial planned for 2023 will do the same, but for patients with treatment-resistant focal epilepsy. Both are proof-of-concept trials to determine whether the approach works and can improve treatment outcomes for these patients. If successful, the team plans to extend their technology to study other aspects of the brain, such as those involved in multiple sclerosis, stroke rehabilitation, depression, and the effects of psychedelics. 

For about a third of epilepsy patients, drugs don’t help. Noninvasive stimulation, in which electrical currents are painlessly delivered to the brain, has been shown to help alleviate the frequency and intensity of seizures. But the technology is still pretty new and needs some refining. This is where a virtual brain could prove useful. 

The digital avatar is essentially a mathematical model running on a computer, says Giulio Ruffini, coordinator of the Neurotwin project and chief science officer and cofounder of Neuroelectrics, a Spanish health tech startup that is developing noninvasive therapies for neurological disorders like epilepsy. To make a digital double for a patient with epilepsy, the Neurotwin team takes about half an hour’s worth of MRI data and about 10 minutes of EEG (electroencephalography) readings and uses these to create a computer model that captures the electrical activity of the brain, as well as to realistically simulate the brain’s main tissues, including the scalp, skull, cerebrospinal fluid, and gray and white matter. 

The twin will include a network of embedded “neural mass models,” says Ruffini. These, he says, are basically computational models of the average behavior of many neurons connected to each other using the patient's “connectome”—a map of the neural connections in the brain. In the case of epilepsy, some areas of the connectome could become overexcited; in the case of, say, stroke, the connectome might be altered. Once the twin has been created, the team can use it to optimize stimulation of the real patient’s brain “because we can run endless simulations on the computer until we find what we need,” Ruffini says. “It is, in this sense, like a weather forecasting computational model.”

For example, to improve treatment for an epilepsy patient, the person would wear a headcap every day for 20 minutes as it delivers transcranial electrical stimulations to their brain. Using the digital twin, Ruffini and his team could optimize the position of stimulating electrodes, as well as the level of current being applied. 

Digital twinning any organ opens up a whole host of ethical questions. For example, would a patient have the right to know—or to refrain from knowing—if, say, their twin predicts that they’ll have a heart attack in two weeks? What happens to the twin after the patient dies? Will it have its own legal or ethical rights? 

On the one hand, virtual body doubles provide us with exciting, revolutionary pathways to develop new treatments, says Matthias Braun, an ethicist at the University of Erlangen-Nürnberg, Germany, who has written about the ethics involved in the use of digital twins in health care. “But, on the other hand, it provides us with challenges,” he continues. For one thing, who should own a digital twin? The company building it? “Or do you have a right to say, well, I refuse the use of specific information or specific predictions with regard to my health insurance or with regard to the use in other contexts? In order to not be an infringement on autonomy or privacy, it is important that this specific person has control of the use [of their digital twin],” he says. Losing that control would result in what Braun dubs “digital slavery.” 

Ana Maiques, the CEO of Neuroelectrics, says the company is already grappling with the issue of what happens to the extremely personal data a digital twin is built upon. “When you're doing these kinds of personalizations, you have to ask difficult questions, right? Who's going to own that data? What are you going to do with data?” she asks. 

The project has enlisted researchers to dissect the ethical and philosophical components of the endeavor, including Manuel Guerrero, a neuroethicist at the University of Uppsala, Sweden. For Neurotwin, a project based out of Europe, the data gathered will be protected by the European Union’s General Data Protection Regulation (GDPR). This means any use of the data requires the consent of its owner, Guerrero says. 

Guerrero and his team are also exploring whether the term “digital twin,” which was first coined for manufacturing, is still the most apt term for copying something as intricate and dynamic as a living brain or heart. Could its use lead to misunderstandings or raised expectations within society? “[The brain] is much more complex than other types of twins that are coming from the manufacturing system, so the notion of a twin for the brain is something that, within the neuroscientific community, is being debated,” he says. 

And taking on the brain is many orders of magnitude more complex than modeling the heart or kidney, in addition to being potentially more ethically complex. “We are creating fairly sophisticated computational models of the brain,” Ruffini says. “At some stage, I think it will become blurry to what extent this digital twin is a digital twin or it's a sentient being.” 

Braun says it’s time to reckon with these thorny questions. “To me, these are really important challenges we have to face now,” he says. “We know what happens if you just say, ‘Well, just develop a technology—and then we'll see,’” he adds, warning of the dangers that come with pushing ethical and moral consequences off to a later date. 

But the Neurotwin team says that, if done right, this digital twinning could dramatically improve both patient outcomes and what we know about tricky-to-treat brain disorders. “We are working to really help people suffering from brain diseases from a completely different perspective,” says Maiques. “We like to call it a new category of therapeutics, where you're really using the power of physics and math to decode the brain.”

Christine Picard’s search for a better bug to feed the world starts with dead bodies. Well, not the corpses themselves, but the blow flies, flesh flies, and other squirmy, wriggly things that wing their way to corpses in the minutes and hours after death. Picard, a forensic entomologist at Indiana University–Purdue University Indianapolis, studies why some insects grow much more quickly than others. This is important for criminal investigations, because the maturity and type of insects found on a body can help nail down exactly when someone died. But Picard’s research on corpse-munching flies is starting to have an effect way beyond autopsy reports: Now her focus is on food.

Don’t worry. It’s (mostly) not for you.

The booming insect farming industry is becoming extremely interested in what makes some larvae grow faster and fatter than others. In 2021, Picard became one of the lead researchers at the Center for Environmental Sustainability Through Insect Farming, a new US-based research center that wants to make farming insects much more efficient. Although the industry is growing, by pure numbers it is still relatively puny: European farms only produced a few thousand tons of insect protein in 2020, a drop in the ocean compared to other sources of protein. These small production numbers keep the price of farmed insects high. One way to solve the problem is to selectively breed fatter and faster-growing bugs—the same approach that the livestock industry has used for centuries to make farm animals more productive.

Breeding better bugs is a high-stakes endeavor. There are now dozens of startups farming insects to sell as ingredients in pet, livestock, and human food. Insects have long been touted as a healthy, low-carbon source of protein that provides an antidote to animal agriculture’s many ills, like methane-burping cows and chicken farms that fuel the growth of antibiotic resistant bacteria. If more of us could join the 2 billion or so people who already eat insects, then we might start putting a dent in the 14.5 percent of all greenhouse gas emissions that currently come from farming livestock.

But you might have noticed that the much-promised future of cricket smoothies for breakfast, cicada sushi for dinner, and mealworm cookies for dessert hasn’t quite materialized yet. Instead, it’s animals—our pets and livestock—that are really driving the edible insect revolution. In Australia, dogs can chow down on pumpkin and mealworm biscuits made by Buggy Bix. In Europe, the brand Tomojo sells all kinds of animal treats supplemented with black soldier fly larvae. Mars, the world’s largest pet food manufacturer, now sells its own brand of insect-supplemented food for cats and dogs. For now, pet food is the largest market for insect protein.

Antoine Hubert, CEO of the French insect-farming startup Ÿnsect, says that the ground mealworms he sells to pet food firms makes up more than 50 percent of his company’s revenue. Pet owners pay a premium to feed their animals food with a lower environmental impact. “It’s where the market adoption has been the biggest,” says Hubert. Although the meat component of pet food typically comes from the bits of animals that humans don’t always eat, it still has a sizable environmental impact. One 2020 study from the University of Edinburgh found that the production of dry pet food alone is responsible for between 1 and 3 percent of global emissions from agriculture.

But feeding pets is only the beginning for insect farming. In 2022, for the first time the EU will allow farmers to feed insects to pigs and poultry—a reversal of rules that banned feed made from animal remains in the wake of the mad cow disease outbreak in the mid-1990s. The poultry and pig feed market is huge: There are 146 million pigs in the EU and 7.2 billion chickens are slaughtered for meat there every year. Even supplying a tiny slice of this market with insect protein would require massively boosting insect production. Ÿnsect is currently building one of the world’s largest insect farms in northern France, but scaling up to feed pigs and poultry would require hundreds of similar farms, Hubert says.

It’ll also require more meatier and faster-growing bugs. One of the major components of chicken feed is soy, which is extremely cheap and widely used across the world. If insect farmers want to start replacing cheap commodity crops like soy, then they’ll need to find a way to bring down costs.

That’s where insect geneticists like Picard come in. “There’s just not enough production right now,” she says. Tuure Parviainen, CEO of Finnish insect farming startup Volare, agrees. “The demand is there, but the production needs to be taken to quite a large scale for the big producers to actually start making a product,” he says. This is just as true for pet food as it is for poultry: The volumes are so huge that big manufacturers aren’t quite ready to go all-in on insects. “The supply is not really there so that they could flip a switch and change ingredients yet,” says Parviainen.

One way to ramp up is to make sure that insect farms are as productive as possible. Scotland-based startup Beta Bugs runs breeding programs to develop more productive versions of the black soldier fly—one of the most commonly farmed insects. “What we have is effectively a very kind of raw material which can then be improved through selective breeding,” says CEO Thomas Farrugia. “Increasingly I think people are starting to realize that this is how we make the industry scale over time.”

Fortunately for insect breeders like Farrugia, time is on their side. Although it might have taken humans thousands of years of breeding to come up with modern cow varieties, insects have much, much shorter life cycles. A black soldier fly is ready to harvest about 14 days after hatching. Its entire life cycle can take around six weeks. “What this means is that you can cram a hell of a lot of selective breeding in a year,” says Farrugia. The trick to breeding a better bug, Farrugia says, is to balance different traits off against each other. You could have one variety of bug that produces lots and lots of skinny larvae, or another that produces a smaller number of fatter young. As larvae mature, the nutrients inside them also change, so one of the tricks of breeding is hitting the sweet spot between fat, juicy bugs and those that are at the right stage in their life cycle.

That said, they don’t want insects that mature too quickly, because the insects are shipped to insect farms while they’re still in the egg stage, to make sure they’re fresh when they arrive. Beta Bugs breeds its black soldier flies in a facility just outside of Edinburgh. From there, eggs are packaged and sent across the EU. Picking the right courier is key, Farrugia says. Black soldier fly eggs hatch in about four days, so if a package is delayed the customer may end up with a delivery that is slightly more alive than they anticipated.

In France, Ÿnsect has launched a breeding program to study the genetics of Tenebrio molitor, the mealworm beetle. The company has already collaborated with the French national center of genome sequencing, Genoscope, to sequence the mealworm’s genome and has also identified a strain of the buffalo worm, a close relative of the mealworm, that grows 25 percent faster than the original strain.

There is a huge amount of genetic diversity inside insects just waiting to be discovered, Picard says. Fruit flies and black soldier flies diverged from each other around 200 million years ago, so the differences between species can be huge. Some are generalists, munching on any kind of waste, while others prefer specialized diets of manure or certain fruits. This might make certain species particularly suited to specific tasks. The oil from mealworms could be a suitable replacement for palm oil, for example, while others might have the specific nutrients that piglets need. We already have different cows for different kinds of beef, Farrugia points out. It might be just a matter of time until we’re chowing down on Angus Aberdeen mealworms, too.

On Christmas, scientists launched the James Webb Space Telescope and sent it about a million miles from Earth. This summer, the technological marvel will begin collecting never-before-seen images of the cosmos. But between now and then, NASA researchers and their European and Canadian colleagues have their work cut out for them.

They have a many-step process to ensure that the powerful, expensive telescope’s instruments are ready to successfully collect data about everything from faint planets to the distant universe. “Everything’s just about on schedule, but we’re busy people for the next six months. There’s an awful lot to do,” says John Mather, JWST senior project scientist at NASA’s Goddard Space Flight Center in Greenbelt, Maryland.

The hardest part might already be done: The spacecraft launched without a hitch, and over the next couple of weeks it delicately unfurled its huge, kite-shaped sun shield, designed to block heat and light from the sun, moon, and Earth, and moved its 18 hexagonal mirror segments into place. “We’re incredibly excited. It was a nail-biter for the first month, and thankfully the deployments went really smoothly,” says Analyn Schneider, the project manager of JWST’s Mid-Infrared Instrument (MIRI) at NASA’s Jet Propulsion Laboratory in Pasadena, California.

All the while, the telescope was traveling to its special parking spot at the L2 Lagrange point, where it balances the gravitational pull of the sun and Earth. (Other spacecraft, including the European Space Agency’s Planck Telescope, have been deployed to the same area.) Keeping a spacecraft in that position is gravitationally unstable, sort of like balancing a ball on an upturned bowl. Webb will regularly drift away from L2, requiring little bursts of fuel every few weeks to nudge it back. But it should have plenty left, because scientists maneuvered the telescope to conserve fuel on its outbound trip. Now, the JWST team expects it to run much longer than its planned 5- to 10-year mission, perhaps lasting as long as its predecessors, the Hubble and Spitzer space telescopes. “The ballpark is probably 20 years of life. It depends on how good we are at steering our unstable car,” Mather says.

Since the spacecraft’s now so far away, Mather, Schneider, and their team have to send and receive radio signals through NASA’s Deep Space Network, an international array of giant antennas managed by JPL. When a programmer inputs a command and waits for an acknowledgment from the spacecraft, that signal could be relayed through an antenna in California’s Mojave Desert or one in Eastern Australia, for example. But there’s a slight delay, because of the distance. “If something bad happens, we won’t know for five seconds,” Mather says. (That’s still pretty quick for space transmissions. For example, messages to our Martian ambassadors like the Perseverance rover involve a delay of five to 20 minutes.)

Now that everything’s in place, the JWST team has begun the “commissioning” process for the instruments, setting up the complex cameras and detectors and making sure they work as they’re supposed to, Schneider says. Last week, they conducted their first tests with the Near-Infrared Camera (NIRCam), allowing the first photons to hit the camera. It’s not actually capturing images yet, but this is a step toward doing so. Eventually, scientists will use NIRCam to discover new planets and glimpse some of the first galaxies.

Once they can take real test images, such as of nearby, previously photographed stars, the first batches will be blurry and out of focus. But that’s normal. Those tests will enable the Webb team to gradually align the telescope and adjust the mirror segments until the images look clear. 

Unlike Hubble’s cameras, which mostly scan the universe at visible-light wavelengths, Webb’s will be sensitive to infrared light, allowing it to probe the early days of the universe and to penetrate dust clouds. But infrared light is essentially heat radiation, so the detectors can’t be contaminated with any other heat, either from the sun or from the spacecraft itself. JWST’s three near-infrared instruments have to be cooled to about -389 Fahrenheit, while MIRI will come within 7 degrees of absolute zero, or about -447 F. Scientists will eventually use MIRI to study the birthplaces of stars. When possible, they’ll use MIRI’s camera and spectrograph, which break down light into its full spectrum of colors, like a rainbow, to look for signs of water, carbon dioxide, and methane; all are common on Earth and might be signs of life-friendly places elsewhere. NIRCam’s detectors can work when they’re slightly warmer than the others, but to function properly, all of the infrared instruments on board have to be cooled down to extremely frigid temperatures.

Because the instruments are behind the sun shield, they will be cooled by space itself—hundreds of degrees colder than anyplace on Earth—while radiating their heat away. For MIRI, engineers designed a special “cryocooler” to chill it down further. “It’s essentially a refrigerator that’s built up with four stages, each stage cooling the next. None of the components in the cryocooler are life-limited. We expect it to continue chugging along as long as we continue to get power from the solar arrays,” says Konstantin Penanen, a cryocooler specialist at JPL.

That’s a major advantage over Spitzer, whose instruments depended on its supply of cryogen, a liquid helium coolant that ran out in 2009. NASA continued to use the space telescope for a few years after that, during the “warm Spitzer mission,” but its mid-infrared detectors were no longer viable.

The JWST team has other challenges ahead, though, like making sure that as the spacecraft cools, little droplets of water vapor escape into space rather than condensing and turning into ice on the mirrors or detectors. (That would make images blurrier.) And over time, tiny micrometeorites, smaller than grains of sand, will likely hit parts of the telescope. But NASA prepared for that too, by making the sun shield five layers thick, so it can withstand those little impacts and minimize the damage.

The spacecraft also depends on some mechanical parts that don’t have a backup. “Webb is so much more complex,” says Sean Carey, who was an astronomer at Caltech’s Spitzer Science Center until it closed last September and is now at NASA’s Exoplanet Science Institute. “It has over 1,000 moving parts. Spitzer had four: the aperture cover, which was ejected once and gone; a focus mechanism, which we moved twice at the beginning of the mission and never moved again; the scan mirror for MIPS; and the shutter for IRAC,” the mid- and near-infrared instruments. If JWST encounters a critical problem, it’s too far away for an astronaut with a screwdriver to be dispatched to repair it, as NASA did for Hubble. 

For now, the next major milestone for the JWST team is to complete the cooling for all the instruments, including MIRI, which will reach its extremely cold temperature range by early April. The long and careful process of aligning the telescope’s mirrors should be done by May. And then it will finally be time for what everyone’s been waiting for: They’ll probably start up the science program—which means actually taking images and data—in June, Mather says.

“I’m thrilled with how well we’re doing," he says. "Nothing has come up that we couldn’t solve.”

The mouse didn't look like much. It had the same red beady eyes and white fur as any other laboratory mouse. Sure, its DNA had been tweaked to make it ideal for testing anti-cancer drugs, but that wasn’t so unusual either. The year was 1988, and it had been more than a decade since researchers at the Salk Institute showed it was possible to create genetically modified mice by inserting viral DNA into mouse embryos. Plenty of other genetically modified animals would be created in the following decades, but none of them would prove as important—or controversial—as OncoMouse.

What made OncoMouse remarkable was its paperwork. On April 12, 1988, the US Patent and Trademark Office issued a patent for it—the first for any living animal. The patent turned a mouse—which had been modified to be more susceptible to cancer—into a legally protected invention, with a patent that prevented anyone else from making or selling mice with the same genetic tweaks. (Or, at least for the 20 or so years that most patents last.) The patent was granted to Harvard University, which passed on the exclusive license to the main funder of its research: DuPont. Soon, the chemical giant was printing T-shirts with an OncoMouse silhouette emblazoned across the chest, and selling researchers the new invention for $50 a mouse.

That patent changed science forever. After OncoMouse, scientists rushed to invent—and patent—other animals that would be useful in their research. Mostly this meant mice, but occasionally other species were patented too, as in the case of rabbits engineered to be susceptible to HIV infection. OncoMouse was used in countless breast cancer studies and helped researchers understand the genetics behind human susceptibility to cancer.

But OncoMouse also raised an awkward question: Where do we draw the line between what belongs to humans and what belongs to nature? And if we could patent only animals that currently exist, what’s to stop us from patenting species that died out long ago? It’s a moral conundrum right out of Jurassic Park, but one that lawyers and scientists are now grappling with for real. Colossal—a startup cofounded by the Harvard geneticist George Church—wants to resurrect a woolly mammoth within the next six years. Its CEO, Ben Lamm, is confident that a mammoth is patentable. But bringing back a species that last stomped the Earth 4,000 years ago raises all kinds of questions that scientists warn we’re not fully prepared for. Can someone really patent a mammoth? And if they can, should they?

Time to contend with these questions is running out. “There is an awful lot of stuff going on at the moment,” says Mike Bruford, a conservation biologist at Cardiff University who helped draft the International Union for Conservation of Nature’s guidelines around de-extinction. Bruford is worried that most de-extinction work is being done by private companies and that scientists can’t be sure of their intentions. “The academic community and the conservation community are, generally speaking, peripheral in this,” he says. When it comes to deciding where—or if—de-extinct animals will be released into the wild, the legal status of those animals will matter an awful lot.

The science of de-extinction—bringing long-extinct species back into existence—is edging closer to possibility. In January 2000, the last living bucardo was killed by a fallen tree. The species of wild goat native to northern Spain had been pushed to the brink of extinction by hunting and was finally done in by a gust of wind.

It was an unceremonious end to a species, but the year before the last bucardo died, scientists in Spain had taken a small chunk of tissue from its ear in the hope that they could eventually use the DNA within to bring the species back to life. That’s exactly what they did in 2003, taking the bucardo’s DNA and putting it inside eggs that were then implanted inside other goat species, close living relatives. Of the 208 embryos, only one made it to term. For a brief moment, the bucardo was back. But then the kid started gasping for breath. Minutes later it was dead—an autopsy later revealed a lung defect common in cloned animals. The bucardo had gone from extinct to extremely endangered and then back to extinct again in a few short minutes.

Other attempts at cloning endangered species have been more successful. In 2020, scientists cloned a black-footed ferret for the first time. That clone—called Elizabeth Ann—is the genetic copy of a wild female called Willa who died in the 1980s. Once widespread throughout the Great Plains in the US, the black-footed ferret was thought to be extinct until a ranch dog helped scientists discover an 18-strong colony in Wyoming in 1981. Although there are now approximately 370 black-footed ferrets in the wild, the species is still extremely endangered, which is why conservation specialists are carefully searching for a mate for Elizabeth Ann.

It's extremely unlikely we’ll ever see a patent for Elizabeth Ann or other animals resurrected through cloning. Most legal systems make it impossible to patent things that occur in nature. You can’t patent an animal or plant simply because you found it first; you need to prove that you’ve invented something. Elizabeth Ann is—legally speaking—an obvious product of nature. Her DNA is a near-exact copy of Willa’s—she’s a duplication, not an invention. The scientists who cloned Dolly the sheep in 1996 hoped to secure a patent, but they were turned down for exactly this reason. “Dolly's genetic identity to her donor parent renders her unpatentable,” wrote a US Court of Appeals judge in 2013, concluding a long legal battle.

But cloning isn’t the only possible path to de-extinction. In September 2021, the startup Colossal launched with the announcement that it had raised $15 million to bring back the woolly mammoth. Although Colossal positions itself as a leader in de-extinction—its website has a whole page dedicated to the term—the startup isn’t exactly resurrecting woolly mammoths. There isn’t a surviving mammoth genome that’s complete enough to implant directly into an egg cell, so cloning is out of the question. What Colossal’s scientists want to do instead is use their knowledge of the mammoth genome to edit the DNA of an Asian elephant so that it more closely resembles that of their ancient, hairier, cousins.

“We’re not de-extincting the mammoth. We’re de-extincting genes to essentially make Asian elephants cold-tolerant,” says Colossal CEO Ben Lamm. The end result would be an elephant-mammoth hybrid that Lamm describes as a “functional mammoth” or an “Arctic elephant.” Eventually Lamm wants to release the Arctic elephants into the Siberian tundra, where he hopes they will help recreate the ancient steppe ecosystem, restore grasslands, and help keep carbon locked in the permafrost. (Whether this would actually happen or not is up for debate.)

Colossal has already sized up a spot for its functional mammoths. Pleistocene Park in the northeastern corner of Russia is a nature reserve maintained by the Russian ecologist Sergey Zimonv and his son, Nikita. The 50-square-mile stretch of tundra is being repopulated with yaks, horses, and bison that the Zimovs hope will uproot and trample away shrubs and trees, making way for the grasslands that covered the area during the Pleistocene Epoch, between 2.6 million and 11,700 years ago. A woolly mammoth—or at least an Asian elephant playing the role—would be the park’s crowning glory.

Despite the nod to Jurassic Park, Lamm says that his goal with Colossal isn’t to directly monetize the mammoths themselves, but to patent and license other technology the company develops along the way. For example, they might need to create giant artificial wombs to grow the mammoth-elephant hybrids, and that technology might help extremely premature human babies survive outside the body. Other techniques they develop for gene editing or storing animal DNA might be helpful for scientific research or conservation efforts. “I think you may get more value out of the technology than the resulting genomes,” Lamm says, although he’s not “closing the door” to patenting whole animals one day.

A project at the nonprofit Revive & Restore, which helped clone the black-footed ferret, is using a similar gene-editing approach to Colossal, but this time to bring back the extinct passenger pigeon. In both of these cases, the aim isn’t to perfectly recreate the extinct species, but to create a hybrid animal that is close enough to the extinct one that it fits into the same ecological niche as its long-dead predecessor. Passenger pigeons may have once been the most numerous birds on the planet, says Ben Novak, a scientist who leads the passenger pigeon project at Revive & Restore. Before they went extinct in 1914, the birds lived in dense flocks across the US and Canada, and their diet of seeds, fruit, and nuts helped build the forests of the northeastern US. Reintroducing the species—or one like it—to the area might help protect these fragile forest ecosystems.

A hybrid approach to de-extinction might be inventive enough to qualify for patent protection. Since mammoth-elephants have never existed in nature, they might not run afoul of the rules that exclude clones from patenting. One recent paper in the Journal of Law and the Biosciences noted that some legal experts are confident that de-extinct species can be patented, at least in the US. (In the European Union, patents can be denied on moral grounds, much to the chagrin of the scientists behind a balding mouse created to test hair loss treatments.) The authors point to a few reasons why companies might want to patent de-extinct animals: to entice investors with the promise of future licensing revenue, to stop other companies from working on the same animals, and to make sure they have exclusive rights to display the animal in a zoo or park.

But Andrew Torrance, a professor of law at the University of Kansas, isn’t so sure that US law will allow it. He points to a legal battle over some patents that would have given a genetic testing firm the exclusive rights to isolate and sequence the human BRCA1 and BRCA2 genes. Mutations in these genes can dramatically increase the risk of breast and ovarian cancer. In 2013, the US Supreme Court ruled that since these two genes occur naturally, they aren’t eligible for patenting. A court may decide that editing an Asian elephant to be more like a mammoth is also recreating something that existed in nature—albeit something that died out thousands of years ago. “Anything you can show was found, or is found in a genome, that will not be patentable in the US and largely unpatentable in other countries as well,” says Torrance, who in 2015 went to Newcastle, England to join a conference discussing the possibility of bringing back the flightless, penguin-like Great Auk, which has been extinct since 1844.

And while we still don’t know if de-extinction will be possible, if it works, it may be profitable. In 2013, the same year that the Supreme Court ruled on the breast cancer genes, three lawyers wrote a lengthy paper arguing that the patent question would have to be addressed sooner or later. De-extinction companies might want exclusive rights to showcase the animals in a purpose-built park, à la Jurassic Park. A company might resurrect the Cuban Macaw or Carolina Parakeet and hawk the birds to parrot fanciers eager to pay a premium for rare birds. Where there is money to be made, the urge to patent is unlikely to be far away. Eventually, Torrance says, the law will have to adapt to these new situations—however far-fetched they seem to us now.

There is, of course, a bigger question: Not just if we can patent revived species, but whether we should. That could end up having a lot to do with what—or whom—de-extinction is for. De-extinction could be seen as a recompense for the hundreds of species humans have partly or wholly driven to extinction. Novak thinks that any species wiped out by humans should be seen as a legitimate candidate for de-extinction—as long as there is still a place left where they can live naturally. For a conservationist like Bruford, the important question is really whether there is a niche in an ecosystem that needs to be filled and whether a resurrected species is the right option there.

Sometimes that niche may have disappeared altogether. It has been thousands of years since woolly mammoths roamed in Siberia, after all. And rather than bringing back extinct species, another way conservationists could fill an ecosystem gap is by introducing a similar species from a different area. For example, Bruford is involved in a project bringing the Aldabra giant tortoise to an island near Mauritius to fill in the gap left behind by the extinct Mauritian tortoise. Others have proposed introducing heat-resistant species of coral to areas endangered by climate change.

If we did bring extinct animals back into modern ecosystems, we may end up running into other serious problems, says Bruford. Mammoths are huge wide-ranging animals that might be hard to contain, and we don’t know if the diseases that may have kept mammoth populations in check still exist today. “It’s not like Jurassic Park, when it’s all on some little fictional island in the middle of the Caribbean. These are big countries with big borders which are porous,” he says.

There’s also the not-insignificant question of how de-extinct animals would be classified. Would a gene-edited Asian elephant be considered a mammoth, an elephant, or something in-between? Would it go immediately on the endangered species list? Or—because it had never existed before—would it technically be an invasive species and prohibited from most areas?

For Novak, although he supports de-extinction, he doesn’t think that the industry should exist for profit, or that a resurrected species should ever be patented. “We are a byproduct of the incredible story of this planet, and it’s an incredible amount of arrogance to believe that we could have some kind of legal right over an entire population of organisms,” he says.

Most of his scientific publications are available online for people to access for free, and those that aren’t he gives away to anyone who asks. If he manages to resurrect the passenger pigeons, Novak says he’ll never sell one. In fact, Revive & Restore had run a mammoth de-extinction project for nine years without attracting enough funding to really get the project underway, Novak says. The nonprofit originally intended to work toward repopulating the tundra in Eurasia and North America with elephant-mammoth hybrids, and its webpage says it brokered the introduction between geneticist George Church and Sergey Zimonv before eventually handing the project over to Colossal.

The revamped, now for-profit project quickly attracted funding from Breyer Capital, Tony Robbins, the Winklevoss brothers, and filmmaker Thomas Tull, whose production firm, incidentally, was behind Jurassic World. “The fact of the matter is that [de-extinction] doesn't attract money. It only attracted money when the idea of profit was brought to the table,” Novak says.

But without private investment, de-extinction might never get off the ground, argues Lamm. “I mean, it’s expensive, from a process perspective,” he says. Colossal will have to raise even more money to keep the project going, and Lamm says that the technologies the startup develops along the way will hopefully benefit health care, research, and conservation. “The de-extinction technology stack can not only be leveraged for species like mammoths, but also for small populations like the northern white rhinos and others,” he says.

Patents—or at least profit—might just be the price conservationists have to pay. And although he vehemently rejects the for-profit de-extinction model, even Novak has an idea he wants to patent. It’s for a genetically modified pigeon that would be much easier to gene-edit than existing birds, and he thinks it could save researchers a lot of time. If his idea works, and he’s granted a patent, he’d like to channel the funds from his invention back into his nonprofit de-extinction work. “We have to make money. The whole world revolves around money,” he says. “So I’d like to try and get a little piece of my pie too.”