https://www.science.org/content/article/breakthrough-2025
2025 BREAKTHROUGH OF THE YEAR
GOOD MORNING, SUNSHINE
The seemingly unstoppable growth of renewable energy is Science’s 2025 Breakthrough of the Year
Since the Industrial Revolution, human society has run on ancient solar energy—captured by plants hundreds of millions of years ago, stored in fossil fuels, and dug and drilled from the earth. But this year momentum shifted unmistakably toward the energy that streams from the Sun today. Renewable energy, most of it from sunlight itself or from wind, ultimately driven by the Sun, overtook conventional energy on multiple fronts.
This year, renewables surpassed coal as a source of electricity worldwide, and solar and wind energy grew fast enough to cover the entire increase in global electricity use from January to June, according to energy think tank Ember. In September, Chinese President Xi Jinping declared at the United Nations that his country will cut its carbon emissions by as much as 10% in a decade, not by using less energy, but by doubling down on wind and solar. And solar panel imports in Africa and South Asia have soared, as people in those regions realized rooftop solar can cheaply power lights, cellphones, and fans. To many, the continued growth of renewables now seems unstoppable—a prospect that has led Science to name the renewable energy surge its 2025 Breakthrough of the Year.
Small rooftop solar systems provide a cheap and reliable source of energy for residents of the Kibera slums of Nairobi, Kenya.DONWILSON ODHIAMBO/SOPA IMAGES/LIGHTROCKET VIA GETTY IMAGES
That promise comes against a backdrop of downbeat news, highlighted at the U.N. climate meeting in Belém, Brazil, in November. Global carbon emissions continue to creep up as countries fall short of cuts pledged in the 2015 Paris climate agreement. The goal of limiting global warming to 1.5°C—always a long shot—now seems completely out of reach. But Hannah Ritchie, a data scientist at the University of Oxford and a climate blogger, is among those who see hope. Thanks to renewables, the long-awaited decline of fossil fuels is in sight, she says. China is “just, just on the cusp … of actually starting to push out coal,” and fossil fuel use in the rest of the world is likely to follow.
China’s mighty industrial engine is the driver. After years of patiently nurturing the sector through subsidies, China now dominates global production of renewable energy technologies. It makes 80% of the world’s solar cells, 70% of its wind turbines, and 70% of its lithium batteries, at prices no competitor can match. “China really mastered this … with the help of the scale of its economy, its manufacturing capacity, and the fierce competition right at home,” says Li Shuo, director of the China Climate Hub at the Asia Society Policy Institute.
As production surged, prices fell and demand took off. Production scaled up to keep pace, further driving down prices and igniting more demand. The result was a virtuous circle in which renewable technologies grew into an industry that now accounts for more than 10% of China’s economy. Wind and solar became the cheapest energy in much of the world.
The tsunami of tech spilling from China’s factories has changed the country’s energy landscape—and its physical one, too. For decades China’s development was synonymous with coal, which produced choking air pollution and massive carbon emissions, still greater than those of all other developed nations combined. Now, solar panels carpet deserts and the high, sunstruck plateau of Tibet, and wind turbines up to 300 meters tall guard coastlines and hilltops (see photo essay, below). China’s solar power generation grew more than 20-fold over the past decade, and its solar and wind farms now have enough capacity to power the entire United States.
China’s burgeoning exports of green tech are transforming the rest of the world, too. Europe is a longtime customer, but countries in the Global South are also rushing to buy China’s solar panels, batteries, and wind turbines, spurred by market forces and a desire for energy independence. In Pakistan, for example, imports of Chinese solar panels grew fivefold from 2022 to ’24 as the Ukraine war pushed up natural gas prices and the cost of grid power. “For people who were asking, ‘How am I going to keep the lights on in my home,’ it was a very obvious choice,” says Lauri Myllyvirta, an analyst at the Centre for Research on Energy and Clean Air. In South Africa, old and unreliable coal plants drove a similar dynamic. Ethiopia has embraced solar and wind amid worries that hydropower, the country’s mainstay, will decline as droughts become more frequent.
Aiming high
Falling prices have propelled a surge in solar and wind energy that far outstrips the growth of any other source. This chart shows installed capacity, a comparison that favors solar and wind, which can produce at full power only a few hours a day, unlike fossil and nuclear. But renewable sources together generated more electricity this year than coal.
*2025 data are full-year estimates. Installed capacity includes grid-connected assets and some off-grid systems.
(GRAPHIC) V. PENNEY/SCIENCE; (DATA) EMBER; BNEF
So far, this is not a story of new technology. China is “more or less relying on the same core [solar] technology that the United States invented half a century ago,” Li says. In those days the U.S. made boutique panels for spacecraft; now, China makes them for the world—better, vastly cheaper, and in staggering quantities.
Technological progress could power future gains. Solar cells today are made of crystalline silicon, but another kind of crystal, perovskites, can be layered in tandem with silicon to make cells that gain efficiency by capturing more colors of light. Material advances are enabling wind turbine blades to get longer and harvest more energy, while designs for floating turbines could vastly expand the offshore areas in which they could be deployed. And the giant lithium-ion batteries now used to store energy when sunshine and wind falter could one day give way to other chemistries. Vanadium flow batteries and sodium batteries could be cheaper; zinc-air batteries could hold far more energy.
In the meantime, climate scientists are already seeing benefits from the existing technology. This year renewables helped bring the growth of greenhouse emissions to a virtual standstill in China and put a global carbon peak within reach. But to meaningfully cut emissions, the world needs to treat the thresholds crossed this year as just a starting point.
“It is worrying to me that China continues to build coal,” says Kelly Sims Gallagher, an energy and climate expert at Tufts University. Dozens of new plants were commissioned in the past year, though many sit idle, waiting to be fired up as “peakers” to meet spikes in demand. In the U.S., President Donald Trump’s administration has declared war on wind and solar development, and cheap Chinese solar panels face formidable trade barriers. U.S. coal consumption is rising this year after a long decline.
The infrastructure needed to take full advantage of wind and solar is another hurdle. China is building battery farms to store green power and long-distance transmission lines to get it to cities, but perhaps not fast enough. Air travel and heavy industries won’t be electrified anytime soon.
But if concerns lurk in the future, looking back brings home the astonishing progress renewables have made. In 2004, it took the world a full year to install 1 gigawatt of solar power capacity. Today, twice that amount goes online each day. Back then renewables had an aura of virtue: Buyers paid a premium over fossil energy because of climate concerns. Now, the real driver is self-interest: lower cost and greater energy security.
That change in motivation may be the most important breakthrough of all, ensuring that this year’s inflection points are just the beginning.
PHOTO ESSAY
Solar panels armor a hillside in China’s Anhui province, parting only for an access road. Distant ridges host wind turbines, another fast-growing component of an energy revolution that has helped ease air pollution and halt the growth of China’s carbon emissions.GEORGE STEINMETZ
THE GREEN GIANT
Images of China’s clean energy infrastructure reveal a transformation of unmatched scale and speed
- BYTIM APPENZELLER
- PHOTOGRAPHY BYGEORGE STEINMETZ
China’s turn to green energy dwarfs any other country’s, as a parade of astonishing numbers attests. In 2024 alone it installed new solar and wind generation equivalent to roughly 100 nuclear power plants, and the pace quickened early this year. Dozens of new, ultrahigh-voltage power lines are marching thousands of kilometers from western deserts where much of the solar energy is generated to the eastern cities where it is used. Hungrily awaiting the bounty of clean energy are millions of electric cars and a sprawling network of high-speed electric trains that can zip between cities 1000 kilometers apart in a morning.
China’s landscape reflects this metamorphosis. The vistas of smog, smokestacks, and coal heaps still exist, but glinting silicon panels now cover hills, deserts, and lakes. One solar farm on the Tibetan Plateau spans more than 400 square kilometers, an area more than twice the size of Washington, D.C. Wind turbines grow ever bigger; one meant for use offshore has blades 150 meters long. Arrays of house-size lithium batteries stockpile excess energy, and more is stored in mountaintop reservoirs, pumped full of water when energy is abundant and tapped as needed by allowing the water to cascade through turbines to a lower lake. The factories that produce the solar panels, turbines, batteries, and cars have added new industrial sprawl—but often without the smokestacks because they are electrified.
These days, the containers in China’s busy ports are packed with new wares: electric cars, solar cells, wind turbine blades. In building up its own green energy system, China has also created an export industry worth nearly $180 billion in 2024, putting low-cost renewable energy within reach for much of the rest of the world (see story, above). The revolution these images document is now going global.
Tipped away from the Sun, a row of mirrors at China’s Shouhang Dunhuang concentrated solar power plant awaits servicing. Unlike solar panels, which generate power on their own, the plant’s 12,000 mirrors reflect sunlight to a central generating station.
At the focus of the Shouhang Dunhuang plant’s mirrors, a tower set aglow by the concentrated sunshine holds molten salt at 565°C. The salt generates steam that powers a turbine—and because the molten material holds heat through the night, the plant can run nonstop.
Solar panels erected in a river’s shallows provide shade for geese as well as power for China’s grid. The country’s bounty of solar power has cut electricity costs and driven investments in batteries and other technologies for storing it at night and on cloudy days.
The fiberglass halves of a single wind turbine blade take shape in a SANY factory in Shaoshan, China. Among the longest in the world, the finished blades extend 107 meters—the length of a soccer field. Demand is so high the factory runs 24/7.
Ready to carve the wind, turbine blades await shipment outside a SANY factory. China produces two-thirds of the world’s wind turbines; most are installed domestically, but international orders are picking up.
A 90-meter needle threading through town, a turbine blade makes its way to a hilltop in Hunan province. China has installed nearly half of the world’s wind power capacity, and although wind is growing more slowly than solar, it generates more reliable power.
A wind turbine is erected in the hills of Hunan province, part of an 18-tower, 88-megawatt wind farm. China installed roughly 50 gigawatts of new wind power in the first half of this year—amounting to some 10,000 turbines this size.
Idling before sunrise in a Nanjing maintenance yard, electric bullet trains wait to be dispatched along the high-speed tracks that span China. At speeds of up to 350 kilometers per hour, the 1300-kilometer journey from Beijing to Shanghai can take just over 4 hours.
Han electric cars get a final inspection at a BYD factory near Shenzhen, China, which can turn out 300,000 cars a year. Han models can travel as far as 700 kilometers on a charge and cost as little as $25,000 in China. Trade barriers have prevented Chinese electric cars from entering the United States.
Cars, most of them electric, await export at the port of Taicang, just upriver from Shanghai. China made more than 12 million electric cars in 2024—70% of global production—and exported 1.25 million of them.
The emerald waters of Qarhan Salt Lake on the Tibetan Plateau beckon tourists—and hold a bounty of lithium, a critical ingredient in batteries for cars and renewable energy storage. Brines in the 160-kilometer-wide lake yielded more than 50,000 tons of lithium carbonate last year, enough for hundreds of thousands of electric cars.
RUNNER-UP
Custom gene editing shows promise for ultrarare diseases
This year, a baby boy with a life-threatening metabolic condition became the world’s first patient to receive a personalized gene-editing treatment. The feat could pave the way for gene editors tailored to people with unique or ultrarare mutations.
KJ Muldoon was born last year in Philadelphia with defects in a gene called CSP1 that encodes an enzyme the liver needs to detoxify ammonia. To avoid a harmful buildup of the compound in the blood and brain, babies with this type of disease must go on a strict protein-limited diet and often need a risky liver transplant.
Soon after KJ’s birth, researchers raced to develop a base editor—a variation of the CRISPR gene-editing tool—that could fix a single spelling mistake in his faulty gene. After testing the bespoke base editor in cells and lab animals, the team won approval to treat KJ in February, when he was 6 months old, using an infusion of lipid nanoparticles to deliver genetic instructions for the base editor. By May, after two more doses, KJ could eat more protein, was gaining weight, and needed less medication to control his ammonia levels.
The researchers now plan to tweak the base editor they used for KJ to treat five patients with similar metabolic disorders caused by other genetic glitches. The U.S. Food and Drug Administration concluded that because these customized treatments will barely differ, they can be tested in a single clinical trial. Such regulatory streamlining could make it feasible to deploy base editing for other rare diseases that until now seemed intractable because they stem from scores or hundreds of different mutations.
A base editor designed to fix a genetic error common to multiple patients also repaired a liver gene in adults for the first time this year. But although such treatments offer fresh hope to many patients, they come with significant caveats, including a hefty price tag and lingering uncertainty about the safety of gene editing.
PHOTO: CHILDREN’S HOSPITAL OF PHILADELPHIA
RUNNER-UP
New weapons against a sexual scourge
Two new drugs for gonorrhea showed their mettle in large clinical trials this year, and both were approved this month by the U.S. Food and Drug Administration. The first new weapons against the sexually transmitted disease in decades, they come at a time when existing treatments are failing.
Gonorrhea, which affects more than 80 million people annually, not only causes pain, genital discharge, and bleeding, but can also lead to serious complications, including pelvic inflammatory disease in women and infertility in both men and women. It increases patients’ risk of HIV infection and can cause blindness if newborn babies’ eyes become infected. Neisseria gonorrhoeae, the culpable bacterium, has developed resistance against nearly every antibiotic doctors have thrown at it. The last class of drugs that still works, cephalosporins, is beginning to fail as well.
In 2019, tests of three candidate replacements showed disappointing results. But in May, The Lancet published the results of a phase 3 trial of gepotidacin, a drug developed by GSK with funding from the U.S. Biomedical Advanced Research and Development Authority. The compound, already approved for urinary tract infections, cured gonorrhea as well as existing drugs. The first in a new class of antibiotics, gepotidacin targets two enzymes, DNA gyrase and topoisomerase IV, that are crucial to DNA replication in bacteria.
The other new drug, zoliflodacin, was developed by Innoviva Specialty Therapeutics in collaboration with the Global Antibiotic R&D Partnership, a Switzerland-based nonprofit. Zoliflodacin also targets DNA gyrase but belongs to another drug class with a different mechanism. A phase 3 study in five countries, published in The Lancet on 11 December, showed it was effective and caused no serious side effects. Both drugs have a major advantage over the most widely used cephalosporin: They can be taken as pills instead of being injected. Scientists hail the arrival of the two compounds, but caution that like every previous drug, they may have a limited shelf life. With a microbe as wily as N. gonorrhoeae, the hunt for new antibiotics never stops.
IMAGE: NOBEASTSOFIERCE/SCIENCE PHOTO LIBRARY
RUNNER-UP
Neurons make a deadly donation to cancer cells
BY MITCH LESLIE
Tumors beguile an assortment of body cells into helping them grow and spread, including neurons. This year, researchers discovered how nerve cells provide this assistance: by passing on mitochondria, the organelles that supply most of cells’ chemical fuel. The result is supercharged cancer cells that more readily spread to other parts of the body, suggesting that blocking the transfers might slow metastasis.
Scientists had previously uncovered evidence that nerves aid cancer. Cutting tumors’ nerve connections often slows their growth or even causes them to shrink, for example. Breaking these links also disrupts cancer cells’ metabolism. To determine why, researchers raised cancer cells and nerve cells together in culture and labeled the neurons’ mitochondria. With a microscope, they spotted nerve cells sending mitochondria to their malignant neighbors across tiny, bridgelike structures. The team also saw evidence of mitochondrial transfers in mice injected with cancer cells and in samples from human prostate tumors, they reported in Nature.
Because cancer cells divide rapidly, they have a hefty appetite for energy. Using cultured cancer cells whose own mitochondria were faulty, the researchers found the hand-me-down organelles provided a metabolic boost. To find out how mitochondrial donations affect cancer growth, they developed a technique that turns cancer cells green once they have acquired the organelles from other cells. The scientists implanted mixtures of nerve and cancer cells into mice and then allowed them to form tumors and metastasize to other tissues. Only 5% of cancer cells in the original tumors turned green, showing they had picked up nerve cell mitochondria, but the proportion rose to 27% in cells from lung metastases and 46% in cells from brain metastases.
Other types of cells also share the organelles, but the new findings suggest neurons may be especially generous—and might eventually help researchers develop treatments that force nerve cells to keep their mitochondria to themselves.
IMAGE: GUSTAVO AYALA AND SIMON GRELET
RUNNER-UP
An all-seeing eye on the sky
BY DANIEL CLERY
A new type of telescope designed to accelerate a new sort of astronomy was completed this year on a mountaintop in Chile. Instead of zooming in on objects of interest, as most other telescopes do, the Vera C. Rubin Observatory will relentlessly sweep across the heavens. Starting early next year, it will record the whole sky in unprecedented detail every 3 days, for 10 years. The bounty will be millions of alerts every night: indications that an object has moved, changed, or suddenly appeared. In 1 year, Rubin will gather more optical data than all other telescopes in history, and it will slowly build up the most detailed 3D map of the cosmos ever assembled, open to everyone via an online portal.
Harvesting starlight so voraciously required groundbreaking technology, including a complex optical system capable of producing distortion-free images over Rubin’s enormous field of view—the size of 45 full Moons—and a car-size camera that can spit out a 3200-megapixel image in seconds. Optical fibers will whisk each image from the summit of Cerro Pachón to California, where banks of computers will analyze it and fire out alerts to astronomers within 1 minute of the image being taken. With 10 million alerts heading their way every night, astronomers will rely on intelligent algorithms to sort the stellar gems from the mundane pebbles.
This deluge of data will touch all areas of astronomy. The observatory will increase the population of known objects in our Solar System many times over—finding, with luck, the postulated Planet 9 lurking beyond Neptune. It will provide ring-side seats to many types of cosmic explosions and help reveal how galaxies form, grow, merge, and evolve into giant agglomerations across cosmic time. It will also aid astronomers studying how unseen dark matter molds galaxies and equally mysterious dark energy boosts the expansion of the universe. Building such a machine is a feat in itself, but few doubt it will be a breakthrough factory in the years to come.
PHOTO: ALIRO PIZARRO DIAZ
RUNNER-UP
Face to face with a Denisovan
BY ANDREW CURRY
This year researchers finally put a face to one of our long-lost relatives—confirming with DNA evidence that a 146,000-year-old skull known as the “Dragon Man” belonged to a Denisovan, an extinct lineage of humans that, like the Neanderthals, once shared the planet with modern humans.
In 2010, geneticists announced they had discovered a new hominin, closely related to Neanderthals and modern humans, based on DNA extracted from a fragment of finger bone found in Denisova Cave in Siberia. But in the 15 years since, Denisovans remained faceless. Small fragments of bone found at other Asian sites from Taiwan to Tibet also yielded Denisovan DNA. But with no complete individuals or even skulls, scientists had no way to know what Denisovans looked like and could not use appearance to identify Denisovan fossils that might already be in museum collections.
That changed this year, after researchers in China managed to extract DNA from an ancient skull found decades ago near Harbin, China. The DNA came from an unusual source—not the teeth or inner ear bones, common sources for ancient genetic material, but a tiny, 0.3-milligram sample of hardened plaque scraped from Dragon Man’s sole remaining tooth. (The genetic material captured in dental plaque comes mostly from bacteria, but also includes DNA from saliva and other fluids in the mouth.)
The fossilized plaque on that blackened molar trapped DNA in a mineral matrix that preserved it longer than porous bone might have. Once sequenced, it linked Dragon Man’s DNA to that of previously sequenced Denisovans. Additional work matched proteins in the plaque to other Denisovan fossils, confirming the skull, with its heavy brow ridges, thick bone, and a powerful jaw, was a Denisovan.
Now that Dragon Man’s identity has been revealed, researchers will have an easier time identifying other Denisovans based on the shape of their bones and teeth. Finding more individuals could help settle an ongoing debate: Were these mysterious hominins a Homo sapiens subspecies or a species of their own?
PHOTO: HEBEI GEO UNIVERSITY
RUNNER-UP
Large language models do science
BY CELINA ZHAO
When Google DeepMind unveiled its protein-structure predictor AlphaFold2 in 2020, it upended expectations for what artificial intelligence (AI) could accomplish in science. Fittingly, it was named Science’s 2021 Breakthrough of the Year—and won a share of the 2024 Nobel Prize in Chemistry for its creators. Few imagined that general-purpose large language models (LLMs), trained on trillions of words and optimized simply to regurgitate humanlike text, might follow suit. That view is now shifting—tectonically—as LLMs scale up. This year, they showcased Ph.D.-caliber acumen across a wide swath of science.
In math, DeepMind used an advanced version of its Gemini LLM to earn a gold medal at the International Mathematical Olympiad, the world’s toughest high school math competition—a feat forecasters in 2021 predicted would remain out of reach until 2043. OpenAI’s GPT-5 also produced original advances for problems in combinatorial number theory and graph theory that had stumped mathematicians for decades.
But LLMs haven’t just aced tests and crunched theory; they’ve also accelerated scientific discovery. In chemistry, a fine-tuned version of Meta’s Llama LLM identified optimal conditions for a previously unreported complex reaction in just 15 experimental runs, saving researchers hundreds of trials that would have taken weeks in the lab. And in biology, Google’s “agentic” AI co-scientist flagged new candidates among existing drugs for liver fibrosis and reproduced in 2 days an insight into the parasitic spread of DNA in bacteria—a result that had taken researchers years to uncover.
Not every experiment has hit the mark. At the ambitious Agents4Science conference, LLMs took charge in formulating hypotheses, analyzing data, and providing the first round of peer reviews. However, the conference left many researchers unconvinced that AI could design and judge scientific questions with adequate rigor.
Yet the leaps made by LLMs have opened the door to a wider AI-for-science gold rush. Tech giants and investors funneled hundreds of millions into spinoffs such as Periodic Labs, Lila Sciences, and OpenAI for Science. And as AI itself begins to improvethe next generation of LLMs, their outer limits are becoming harder and harder to predict.
PHOTO: LAURENT GRANDGUILLOT/REA/REDUX
RUNNER-UP
Triumph of calculation helps resolve particle mystery
BY ADRIAN CHO
For decades, particle physicists have longed for something—anything—their prevailing theory, the standard model, cannot explain. In June, perhaps the most tantalizing sign of a new mystery vanished when a long-running experiment reported that, contrary to its earlier claims, a particle called the muon was not more magnetic than the standard model predicts. Behind the disappointment lurks a triumph: Theorists were finally able to calculate the muon’s magnetism precisely from scratch using a technique called lattice gauge theory.
The muon is a heavier, unstable cousin of the electron. Its magnetism gets a small boost, denoted g-2, from particles that, because of quantum uncertainty, flit in and out of the vacuum surrounding the muon. If those “virtual” particles include ones not found in the standard model, the muon’s magnetism could differ from the theory’s predictions. Beginning in 2001, a U.S. experiment called Muon g-2 suggested the muon’s magnetism was stronger than predicted by about 4 parts per billion.
Making that prediction was never easy. Accounting for the effects of particles called quarks and gluons was a particular challenge because they interact through the so-called strong force, which is nearly intractable mathematically. In 2020, a collaboration called the Muon g-2 Theory Initiative extrapolated from data taken at particle colliders to estimate how much virtual quarks and gluons contribute to the muon’s magnetism. But the best available data still had inconsistencies.
Theorists can also calculate the quarks and gluons’ contribution from scratch, using supercomputers and lattice gauge theory, a numerical technique that simplifies the problem by dividing continuous spacetime into a 4D lattice of points. Thanks to ever-increasing computer power and myriad advances in technique, recent lattice calculations predict the muon’s magnetismwith a precision rivaling that of the data-driven method. In May, the theory initiative revised its prediction by discarding the data-driven method in favor of the lattice method. The new prediction for the muon’s magnetism is higher than previously calculated and agrees with Muon g-2’s final measurement, which was announced shortly afterward. The agreement marks a triumph for lattice gauge theory—even as it scotches a hint of new physics.
PHOTO: RYAN POSTEL/FERMILAB
RUNNER-UP
Xenotransplants set new records
BY JON COHEN
For the past century, hype and sometimes dubious science have marred attempts to heal humans with transplanted animal organs. But “xenotransplantation,” a potential solution to the dire shortage of donated human organs, took impressive steps forward this year thanks to pigs genetically engineered to make their tissues safer for transplants and less likely to suffer rejection from human immune systems.
Most notably, a pig kidney with 69 altered genes worked for nearly 9 months in a New Hampshire man until it failed in October, just a few days short of the longest previous xenotransplant. (That record had been set in 1964 with a natural kidney from a chimpanzee, a species no longer deemed an ethical option for donor organs.) A pig kidney with a mere six modified genes worked nearly as long in a woman in China.
Those efforts extend the previous record for an engineered pig kidney, set with an organ with 10 genetic changes that helped a woman for 4 months before failing in February. In each case, the organs came from pigs modified and bred by companies hoping xenotransplantation can be a profitable business. Two of those companies this year received permission from the U.S. Food and Drug Administration to stage the first bona fide clinical trials of the strategy, a necessary step for regulatory approval.
The donor pigs still need additional, yet-to-be-determined genetic modifications to increase the survival time of transplanted organs, researchers widely agree. They’re also attempting to develop safer, more effective drugs to suppress rejection. And some are trying to devise novel strategies to promote immune tolerance—such as transplanting a pig’s thymus along with its kidney—that do away with the need for immunosuppressive drugs altogether. But this year’s successes have arguably moved xenotransplantation closer to what far too many sensational headlines have celebrated prematurely for decades.
PHOTO: JOE CARROTTA/NYU LANGONE HEALTH
RUNNER-UP
Rice that beats the heat
Crops can tolerate the scorching sunshine of a heat wave if they have enough water, but sweltering nights can spell particularly serious trouble, ramping up respiration, a metabolic process that normally holds steady in the dark. This year, researchers in China discovered a gene that helps protect rice from two impacts of heat: lower yield and poor-quality grain. If bred or engineered into commercial varieties, the gene could help safeguard rice harvests as climate change heats up the fields.
Over the past decade, the team grew 533 varieties of rice in parts of China that have become unusually hot. By crossbreeding the types of rice that performed best, they identified a key gene on chromosome 12, which they called QT12, for “quality-thermotolerant.” Varieties that did poorly in the heat had a variant, or allele, of QT12that activated and caused starch molecules to misalign. The result was chalky, brittle grain with an unappetizing pasty taste, the team reported in April in Cell. But varieties with a QT12 allele that resists turning on when temperatures rise maintained good-quality grain. The allele also protected yield, although it is not clear how.
When the researchers bred the protective allele into a commercial variety of rice called Huazhan, it produced up to 78% more rice and a lower proportion of chalky grain than the current version of Huazhan when exposed to high heat. They also showed that disabling QT12 with gene editing had similar benefits in a research variety of rice.
Varieties of the japonica subspecies of rice, which are grown in cooler regions and fare poorly in heat, might benefit in particular from the protective allele of QT12. Japonicarice doesn’t carry that allele, the researchers found, but it could be added by conventional breeding. They suggest other grain crops, such as wheat and maize, could be equipped with a similar gene to guard them from rising temperatures to come.
PHOTO: NHAC NGUYEN/AFP VIA GETTY IMAGES
BREAKDOWNS
Trump roils U.S. science
Protesters gathered at a Stand Up for Science rally in San Francisco in March.GABRIELLE LURIE/SAN FRANCISCO CHRONICLE VIA GETTY IMAGES
President Donald Trump began to rewrite the playbook on how the U.S. government interacts with the research community within minutes of taking office on 20 January. Seemingly every week since then he has added a page, shattering decades-old norms that helped make the United States a scientific superpower.
The now-defunct Department of Government Efficiency (DOGE), led by billionaire Elon Musk, played a prominent role in the first few months. It forced the National Institutes of Health (NIH) and the National Science Foundation (NSF) to cancel thousands of existing projects, in particular those deemed to involve diversity, equity, and inclusion; climate change; or gender identity and sexual orientation.
Trump also used DOGE to spearhead another prong of the attack: shrinking the size of the federal workforce. Thousands of employees at science agencies were laid off, accepted buyout offers, quit, or retired sooner than planned, resulting in a significant loss of scientific experience. Many of those who remained are managing narrower portfolios after their agency pulled the plug on scores of programs the White House decided—suddenly, and without explanation—were “no longer aligned with the administration’s priorities.” NSF and the Department of Energy’s Office of Science have gone even further, revamping units to correspond more closely to those priorities.
Several of Trump’s choices to lead federal science agencies have triggered fierce opposition. Foremost among them is Secretary of Health and Human Services Robert F. Kennedy Jr., a lawyer and longtime antivaccine activist. Kennedy has decimated senior management at NIH, the Food and Drug Administration, and the Centers for Disease Control and Prevention, along with a vast network of advisory committees. The new heads of those agencies share some of Kennedy’s contrarian views on public health and have altered many existing policies. NIH, for example, added criteria that could politicize its choices about what grant applications to fund.
Claiming to find evidence of antisemitism, the Trump administration froze all federal grants to a handful of top-tier colleges and universities and threatened similar action at scores more. Some agreed to pay sizable settlements and allow greater federal oversight in return for lifting those freezes, while negotiations continue with other schools, notably Harvard University and the University of California, Los Angeles.
An aggressive campaign to detain and deport people alleged to lack valid immigration status has made graduate students on valid visas fearful that they may be detained after traveling abroad, and led to fewer new international students at most U.S. colleges and universities. In response, many institutions have reduced the size of their doctoral programs in science and engineering.
Research advocates have fought many Trump directives in court, with mixed success. Arguably the community’s biggest win to date was blocking attempts by four science agencies to slash payments to universities for so-called indirect costs—money spent on infrastructure and overhead that supports federally funded research. Another silver lining in this gloomy year was that both NSF and NIH managed to spend all the money Congress had given them this year despite the many disruptions to their grantmaking. And Congress appears likely to resist Trump’s request for sharp cuts in the 2026 budgets of most science agencies. Equally important, most of this year’s onslaught has been carried out not through permanent legislation, but through temporary executive orders that Trump’s successor could rescind.
Global health in crisis
A mother cries after her infant son was diagnosed with severe malaria at a hospital in Guinea, where the freeze on U.S. aid has disrupted efforts to control the disease.MARTA MOREIRAS
For decades, aid for global health had solid support across the political spectrum in many countries. But this year marked a dramatic turnaround as wealthy countries slashed billions in funding for organizations that fight disease and malnutrition in other parts of the world.
In January, after President Donald Trump took office, the United States froze all foreign aid, including more than $10 billion for global health, and dismantled the U.S. Agency for International Development. The moves halted deliveries of food and drugs and triggered thousands of layoffs in the U.S. and abroad, although some of the funds were reinstated later in the year. Aid budgets in France, Germany, the United Kingdom, the Netherlands, and Finland have tumbled as well, in part to make room for increased defense spending.
A July report from the Institute for Health Metrics and Evaluation estimated that donations for global health would fall to $39.1 billion this year—a 21% drop. Groups such as the Global Fund to Fight Aids, Tuberculosis and Malaria; the Joint United Nations Programme on HIV/AIDS (UNAIDS); and Gavi, the Vaccine Alliance all have to tighten their belts as a result. The World Health Organization, in which the U.S. has canceled its membership, has been hardest hit proportionally and is going through a painful downsizing.
The cuts are already noticeable on the ground, and will get worse. In July, the Global Fund slashed existing grants for all three diseases it helps fight, forcing countries to scale back spending on commodities such as medicines, diagnostics, insecticides, and bed nets. UNAIDS has reported severe disruptions in HIV prevention and testing in many countries.
Studies modeling the effects of the recent cuts predict a surge in disease cases and deaths, especially in sub-Saharan Africa. Some countries may mitigate the toll by shouldering more of the burden themselves, however, and new donors may step up. But because many poor countries don’t always record causes of death, the precise toll of the shifting political winds may never be known.
AI and fraud degrade the literature
This year several studies exposed how paper mills, artificial intelligence (AI), and research fraud are polluting the scholarly literature with trivial, shoddy, and nonsensical papers—and the pace is accelerating.
An unusually thorough analysis of this long-standing issue found that the number of fake papers—though still relatively small—seems to be increasing far faster than the scientific literature overall. At fault is a complex ecosystem of authors, editors, publishers, and others conspiring to pad authors’ CVs—and the pockets of editors and publishers who profit from article-processing fees for the fake papers and sometimes even take bribes.
The continued rise of AI is also shaping the literature. A massive, cross-disciplinary look at how often scientists turn to AI to help write their manuscripts found steady increases since 2022. Some uses may be legitimate—for instance, if they are disclosed to the journal upon submission and align with the journal’s policy. But disclosure may be the exception. The American Association for Cancer Research recently found that four times as many authors use AI as admit to itacross the 10 journals it publishes.
The transgressions may go beyond nondisclosure. One study reported that unscrupulous actors were using publicly available health data sets—and likely generative AI tools—to churn out trivial papers exploring correlations between nearly every possible combination of condition, contributing factor, and population. Authors are also using AI to concoct letters to the editor and other simple commentaries that have little substance and are simply intended to inflate the author’s publication record, another team reported.
Some publishers are beginning to take aggressive countermeasures. In July, Taylor & Francis paused submissions to its journal Bioengineeredso editors could investigate some 1000 paperssuspected of containing manipulated results or coming from paper mills. And in September, PLOS and Frontiers announced they had begun to automatically reject the vast majority of papersbased on those exploited public health data sets. But the arms race between publishers and AI-fueled malfeasance is sure to continue to escalate.
MULTIMEDIA
18 DEC 2025
BY
- SARAH CRESPI, MEAGAN CANTWELL, GREG MILLER, DAVID GRIMM
This year’s biggest breakthrough and top news stories
On this week’s show: Science names this year’s most significant advance, and we survey the best from our online news section
doi: 10.1126/science.z57oqnf
RELEVANT TAGS:
Last edited by @suen 2025-12-19T10:20:49Z