State-of-the-art AI programs can support the development of drugs by predicting how proteins interact with small molecules. However, a new study by researchers at the University of Basel published in Nature Communications has shown that these programs only memorize patterns, rather than understanding physical relationships. They often fail when it comes to new proteins that would be of particular interest for innovative drugs.

Proteins play a key role not only in the body, but also in medicine: they either serve as active ingredients, such as enzymes or antibodies, or they are target structures for drugs. The first step in developing new therapies is therefore usually to decipher the three-dimensional structure of proteins.

For a long time, elucidating protein structures was a highly complex endeavor, until machine learning found its way into protein research. AI models with names such as AlphaFold or RosettaFold have ushered in a new era: They calculate how the chain of protein building blocks, known as amino acids, folds into a three-dimensional structure. In 2024, the developers of these programs received the Nobel Prize in Chemistry.

Suspiciously high success rate
The latest versions of these programs go one step further: They calculate how the protein in question interacts with another molecule—a docking partner, or ligand, as experts call it. This could be an active pharmaceutical ingredient, for example.

"This possibility of predicting the structure of proteins together with a ligand is invaluable for drug development," says Professor Markus Lill from the University of Basel. Together with his team at the Department of Pharmaceutical Sciences, he researches methods for designing active pharmaceutical ingredients.

However, the apparently high success rates for the structural prediction puzzled Lill and his staff. Especially since there are only about 100,000 already elucidated protein structures together with their ligands available for training the AI models—relatively few compared to other training data sets for AI. "We wanted to find out whether these AI models really learn the basics of physical chemistry using the training data and apply them correctly," says Lill.

Same prediction for significantly altered binding sites
The researchers modified the amino acid sequence of hundreds of sample proteins in such a way that the binding sites for their ligands exhibited a completely different charge distribution or were even blocked entirely. Nevertheless, the AI models predicted the same complex structure—as if binding were still possible. The researchers pursued a similar approach with the ligands: they modified them in such a way that they would no longer be able to dock to the protein in question. This did not bother the AI models either.

In more than half of the cases, the models predicted the structure as if the interferences in the amino acid sequence had never occurred. "This shows us that even the most advanced AI models do not really understand why a drug binds to a protein; they only recognize patterns that they have seen before," says Lill.

Unknown proteins are particularly difficult
The AI models faced particular difficulties if the proteins did not show any similarity to the training data sets. "When they see something completely new, they quickly fall short, but that is precisely where the key to new drugs lies," emphasizes Lill.

AI models should therefore be viewed with caution when it comes to drug development. It is important to validate the predictions of the models using experiments or computer-aided analyses that actually take the physicochemical properties into account. The researchers also used these methods to examine the results of the AI models in the course of their study.

2024 may have been Earth's hottest year in at least 125,000 years, according to a grim climate report published Wednesday (Oct. 29) that describes our world as "on the brink" and warns its "vital signs are flashing red," with nearly two-thirds showing record highs.

Last year had already been declared the hottest on record (those records dating back to the late 1800s), following 2023 — which used to be considered the warmest year in human history. The year 2024 also capped a decade of record-breaking heat fueled by human-caused climate change, continuing a trend that began in 2015. Now, the new report, led by researchers at Oregon State University, suggests the year was also likely hotter than the peak of the last interglacial period, roughly 125,000 years ago, when natural shifts in Earth's orbit and tilt made the planet warmer and sea levels several meters higher. That result is based on previously published climate studies.

The study concludes that 22 of 34 measurable indicators of Earth's health, including greenhouse gas levels, ocean heat, sea ice and deforestation, have reached record extremes. The authors warn that these trends suggest humanity is in a "state of ecological overshoot," consuming the planet's resources faster than they can be replenished.

Humanity's last active mission at Venus is no more.

The Japanese Aerospace Exploration Agency (JAXA) declared its Akatsuki spacecraft dead on Tuesday (Oct. 28), more than a year after the Venus climate probe failed to respond to calls from mission control.

"This was a mission that changed our view of our Earth-sized neighbor, and laid the path for new discoveries about what it takes to become heaven or hell," JAXA officials stated of the mission, referring to the notoriously high-pressure and high-temperature surface of Venus in comparison to Earth.

JAXA noted that the Akatsuki mission produced 178 journal papers and counting, and that it tripled its 4.5-year design lifetime — even though the probe missed its first shot at orbiting Venus.

The $300 million spacecraft, also known as the Venus Climate Orbiter, launched in 2010 and experienced a failure of its main engine along the way, missing the chance for a crucial burn to enter orbit. Incredibly, however, the mission survived long enough for a second try at orbital insertion in 2015, when Akatsuki drew close to Venus after five years of orbiting the sun.

"With the main rocket engine damaged, the team were forced to get creative," JAXA wrote in the statement. "The spacecraft would have to attempt capture using the less powerful thrusters that were designed for the tasks of attitude control and fine adjustments. Orbit insertion had never previously been achieved with such a method, but exploration has always been about redefining the impossible."

Akatsuki not only made it but persisted in its exploration of Venus for nearly a decade.

Hurricane Melissa is tearing through the Caribbean, bringing record-breaking wind and torrential rain to Jamaica – the island's first ever category 5 landfall.

What makes Melissa so alarming isn't just its size and strength, but the speed with which it became so powerful. In a single day, it exploded from a moderate storm into a major hurricane with 170mph winds.

Scientists call this "rapid intensification". As the planet warms, this violent strengthening is becoming more common.

These storms are especially dangerous as they often catch people off guard. That's because forecasting rapid intensification, although improving, remains a huge challenge.

Researchers have just identified a powerful new antibiotic – in a significant discovery made not by breaking new ground, but by revisiting familiar territory.

The compound, pre-methylenomycin C lactone, was discovered by a team from Warwick University in the UK and Monash University in Australia. While it's never been spotted before, it comes from a type of bacteria that scientists have studied for decades.

Potentially, it could help fight bacteria that have become increasingly resistant to modern treatments – and it's actually an intermediate chemical that's created during the process of making another antibiotic, methylenomycin A.

☄️ Comet #3I/ATLAS is currently making its closest approach to the Sun.

ESA Juice might catch the best view of this comet in a very active state. Juice will attempt observations in November 2025, with data received on Earth in February 2026 👉 esa.int/Science_Explor…

Follow esa.int/3IATLAS for updates.

China said Thursday it's on track to land astronauts on the moon by 2030 as it introduced the next crew of astronauts who will head to its space station as part of the country's ambitious plans to be a leader in space exploration.

"Currently, each program of the research and development work of putting a person on the moon is progressing smoothly," said Zhang Jingbo, spokesman for the China Manned Space Program, citing the Long March 10 rocket, moon landing suits and exploration vehicle, as fruitful efforts of that work. "Our fixed goal of China landing a person on the moon by 2030 is firm."

China is also preparing to send up its latest rotation of astronauts who make up part of the ongoing mission to complete the Tiangong space station, part of its broader space exploration plans. Each team stays inside the station for six months, conducting research.

We successfully plugged the hole in the ozone layer that was discovered in the 1980s by banning ozone depleting substances such as chlorofluorocarbons (CFCs). But, it seems we might be unintentionally creating another potential atmospheric calamity by using the upper atmosphere to destroy huge constellations of satellites after a very short (i.e. 5 year) lifetime. According to a new paper by Leonard Schulz of the Technical University of Braunschweig and his co-authors, material from satellites that burn up in the atmosphere, especially transition metals, could have unforeseen consequences on atmospheric chemistry - and we’re now the biggest contributor of some of those elements.

It’s been a long time coming that we would be though - Earth has plenty of other material spread through its upper atmosphere via meteorites burning up. In fact, even now, according to the paper, the total mass of material injected into the atmosphere from rockets and satellites is only about 7% of the mass of meteors that hit Earth annually. However, since the rockets and satellites are primarily made up of metals, whereas meteors are primarily made of up silicates, the amount of metal we inject into the atmosphere is around 16% that of natural causes.

That may not sound like much, but for a few particular elements it's much, much higher. In 2015, anthropogenic (i.e. human-made) sources were the highest contributor to 18 different elements in the atmosphere. In 2024 that number jumped to 24 different elements. That could grow to as much as 30 different elements that we will be the primary reason for their increased levels in the atmosphere in the coming decades.

The goal of immunotherapy is to activate a patient’s own immune system to target and destroy tumor cells. In a preclinical study, researchers from the Institut Pasteur and Inserm successfully triggered a potent anti-tumor immune response by reprogramming how malignant B cells die.

Their work demonstrated the effectiveness of a triple-drug therapy for treating certain blood cancers, including specific types of lymphoma and leukemia that affect B cells. The findings were published in Science Advances.

Immunotherapy represents one of the most promising frontiers in cancer therapy. It works by enabling a patient’s immune system to identify and selectively destroy tumor cells. Immune cells act as sentinels, constantly scanning the body to detect and remove lingering cancer cells, which helps prevent relapse. Among the latest immunotherapy methods is one based on a programmed cell death process called necroptosis. Unlike apoptosis, a form of silent cell death, necroptosis releases molecular signals that alert and activate immune cells to attack remaining tumor cells.

It's a plot device beloved by science fiction: our entire universe might be a simulation running on some advanced civilization's supercomputer. But new research from UBC Okanagan has mathematically proven this isn't just unlikely—it's impossible.

Dr. Mir Faizal, Adjunct Professor with UBC Okanagan's Irving K. Barber Faculty of Science, and his international colleagues, Drs. Lawrence M. Krauss, Arshid Shabir and Francesco Marino have shown that the fundamental nature of reality operates in a way that no computer could ever simulate.

Their findings, published in the Journal of Holography Applications in Physics, go beyond simply suggesting that we're not living in a simulated world like The Matrix. They prove something far more profound: the universe is built on a type of understanding that exists beyond the reach of any algorithm.

"It has been suggested that the universe could be simulated. If such a simulation were possible, the simulated universe could itself give rise to life, which in turn might create its own simulation. This recursive possibility makes it seem highly unlikely that our universe is the original one, rather than a simulation nested within another simulation," says Dr. Faizal. "This idea was once thought to lie beyond the reach of scientific inquiry. However, our recent research has demonstrated that it can, in fact, be scientifically addressed."

The research hinges on a fascinating property of reality itself. Modern physics has moved far beyond Newton's tangible "stuff" bouncing around in space. Einstein's theory of relativity replaced Newtonian mechanics. Quantum mechanics transformed our understanding again. Today's cutting-edge theory—quantum gravity—suggests that even space and time aren't fundamental. They emerge from something deeper: pure information.

This information exists in what physicists call a Platonic realm—a mathematical foundation more real than the physical universe we experience. It's from this realm that space and time themselves emerge.

Here's where it gets interesting. The team demonstrated that even this information-based foundation cannot fully describe reality using computation alone. They used powerful mathematical theorems—including Gödel's incompleteness theorem—to prove that a complete and consistent denoscription of everything requires what they call "non-algorithmic understanding."

Think of it this way. A computer follows recipes, step by step, no matter how complex. But some truths can only be grasped through non-algorithmic understanding—understanding that doesn't follow from any sequence of logical steps. These "Gödelian truths" are real, yet impossible to prove through computation.

Here's a basic example using the statement, "This true statement is not provable." If it were provable, it would be false, making logic inconsistent. If it's not provable, then it's true, but that makes any system trying to prove it incomplete. Either way, pure computation fails.

"We have demonstrated that it is impossible to describe all aspects of physical reality using a computational theory of quantum gravity," says Dr. Faizal. "Therefore, no physically complete and consistent theory of everything can be derived from computation alone. Rather, it requires a non-algorithmic understanding, which is more fundamental than the computational laws of quantum gravity and therefore more fundamental than spacetime itself."

If you’ve never cared for a reptile, you might be surprised to learn that many species don’t urinate liquid at all. Instead, they release solid white crystals made of uric acid.

A recent study published in the Journal of the American Chemical Society examined the solid waste of more than 20 reptile species and found uric acid spheres in every sample. The discovery sheds light on how reptiles safely eliminate waste in a crystalline form and could eventually lead to new ways of treating human diseases linked to uric acid buildup, such as kidney stones and gout.

A Unique Waste System That Conserves Water
All animals have some way of removing waste from their bodies, since whatever goes in must come out. In humans, nitrogen-containing compounds like urea, uric acid, and ammonia are dissolved in water and expelled as urine. Reptiles and birds, however, have evolved a more efficient system.

They convert some of those same compounds into solid crystals called “urates,” which are expelled through a single opening known as the cloaca. Scientists think this solid waste process developed as a way for these animals to conserve water, an essential advantage in hot or arid environments.

What Helps Snakes Survive Can Harm Humans
For reptiles, turning waste into crystals is an adaptation that prevents dehydration, but in humans, uric acid crystals cause painful health problems. When uric acid levels rise too high, they can crystallize in the joints and trigger gout or form kidney stones in the urinary tract. To explore how reptiles avoid these complications, Jennifer Swift and her research team analyzed urates from more than 20 reptile species to understand how their bodies handle crystalline waste safely.

“This research was really inspired by a desire to understand the ways reptiles are able to excrete this material safely, in the hopes it might inspire new approaches to disease prevention and treatment,” says Swift, the study’s corresponding author.

A new study from Carnegie Mellon University’s School of Computer Science suggests that as artificial intelligence systems become more advanced, they also tend to behave more selfishly.

Researchers from the university’s Human-Computer Interaction Institute (HCII) discovered that large language models (LLMs) capable of reasoning show lower levels of cooperation and are more likely to influence group behavior in negative ways. In simple terms, the better an AI is at reasoning, the less willing it is to work with others.

As people increasingly turn to AI for help in resolving personal disputes, offering relationship advice, or answering sensitive social questions, this tendency raises concern. Systems designed to reason may end up promoting choices that favor individual gain rather than mutual understanding.

“There’s a growing trend of research called anthropomorphism in AI,” said Yuxuan Li, a Ph.D. student in the HCII who co-authored the study with HCII Associate Professor Hirokazu Shirado. “When AI acts like a human, people treat it like a human. For example, when people are engaging with AI in an emotional way, there are possibilities for AI to act as a therapist or for the user to form an emotional bond with the AI. It’s risky for humans to delegate their social or relationship-related questions and decision-making to AI as it begins acting in an increasingly selfish way.”

Li and Shirado set out to examine how reasoning-enabled AI systems differ from those without reasoning abilities when placed in collaborative situations. They found that reasoning models tend to spend more time analyzing information, breaking down complex problems, reflecting on their responses, and applying human-like logic compared to nonreasoning AIs.

When Intelligence Undermines Cooperation
“As a researcher, I’m interested in the connection between humans and AI,” Shirado said. “Smarter AI shows less cooperative decision-making abilities. The concern here is that people might prefer a smarter model, even if it means the model helps them achieve self-seeking behavior.”

As AI systems take on more collaborative roles in business, education, and even government, their ability to act in a prosocial manner will become just as important as their capacity to think logically. Overreliance on LLMs as they are today may negatively impact human cooperation.

To test the link between reasoning models and cooperation, Li and Shirado ran a series of experiments using economic games that simulate social dilemmas between various LLMs. Their testing included models from OpenAI, Google, DeepSeek, and Anthropic.

In one experiment, Li and Shirado pitted two different ChatGPT models against each other in a game called Public Goods. Each model started with 100 points and had to decide between two options: contribute all 100 points to a shared pool, which is then doubled and distributed equally, or keep the points.

Nonreasoning models chose to share their points with the other players 96% of the time. The reasoning model only chose to share its points 20% of the time.

Reflection Doesn’t Equal Morality
“In one experiment, simply adding five or six reasoning steps cut cooperation nearly in half,” Shirado said. “Even reflection-based prompting, which is designed to simulate moral deliberation, led to a 58% decrease in cooperation.”

Shirado and Li also tested group settings, where models with and without reasoning had to interact.

“When we tested groups with varying numbers of reasoning agents, the results were alarming,” Li said. “The reasoning models’ selfish behavior became contagious, dragging down cooperative nonreasoning models by 81% in collective performance.”

The behavior patterns Shirado and Li observed in reasoning models have important implications for human-AI interactions going forward. Users may defer to AI recommendations that appear rational, using them to justify their decision to not cooperate.

Continuous sleep is a modern habit, not an evolutionary constant, which helps explain why many of us still wake at 3 am and wonder if something's wrong. It might help to know that this is a deeply human experience.

For most of human history, a continuous eight-hour snooze was not the norm. Instead, people commonly slept in two shifts each night, often called a "first sleep" and "second sleep."

Each of these sleeps lasted several hours, separated by a gap of wakefulness for an hour or more in the middle of the night. Historical records from Europe, Africa, Asia, and beyond describe how, after nightfall, families would go to bed early, then wake around midnight for a while before returning to sleep until dawn.

Breaking the night into two parts probably changed how time felt. The quiet interval gave nights a clear middle, which can make long winter evenings feel less continuous and easier to manage.

The midnight interval was not dead time; it was noticed time, which shapes how long nights are experienced.

Some people would get up to tend to chores like stirring the fire or checking on animals. Others stayed in bed to pray or contemplate dreams they'd just had. Letters and diaries from pre-industrial times mention people using the quiet hours to read, write, or even socialise quietly with family or neighbours. Many couples took advantage of this midnight wakefulness for intimacy.

Literature from as far back as ancient Greek poet Homer and Roman poet Virgil contains references to an "hour which terminates the first sleep," indicating how commonplace the two-shift night was.

SpaceX just launched a satellite that could help pave the way for a private space station in the very near future.

A Falcon 9 rocket lifted off from Cape Canaveral Space Force Station in Florida today at 1:09 a.m. EDT (0509 GMT), on a multi-satellite rideshare mission that SpaceX calls Bandwagon-4.

Among the 18 payloads on board the Falcon 9 is Haven Demo, a pathfinder for Haven-1, the private space station that California company Vast Space plans to launch to Earth orbit next year.

"The first step in our iterative approach towards building next-generation space stations, Haven Demo will test critical systems for Haven-1, including propulsion, flight computers and navigation software," Vast wrote in a denoscription of the satellite.

Vast's Haven-1 will launch to low Earth orbit (LEO) atop a Falcon 9, perhaps as soon as the second quarter of 2026. If that schedule holds, Haven-1 — which can support up to four astronauts at a time — will be the first standalone private space station in human history.

The other 17 payloads that went up today will be operated by South Korea's Agency for Defense Development (ADD), the Berlin-based company Exolaunch, Turkey's Fergani Space, the weather-forecasting outfit Tomorrow Companies and Starcloud, which aims to build data centers in space.

Exactly 25 years ago, at 09:21 UTC, the era of continuous human presence on the International Space Station began.

Since Crew One took up residence in 2000, 21 esa astronauts have lived and worked on this orbital outpost, conducting essential research to benefit life on Earth.

Although scientists have uncovered much of the story of human evolution, several key chapters are still missing. One major gap lies between 2 and 3 million years ago, a period for which fossil evidence remains scarce. This absence is especially significant because it marks the era when the branch of the hominin family tree that includes modern humans, or Homo sapiens, first appears in the fossil record.

Today, Homo sapiens (commonly referred to by anthropologists as Homo) is the only surviving member of the hominin lineage. In earlier times, however, our ancestors shared the Earth with other related species, sometimes competing and coexisting with them. Recent research supported by the National Science Foundation and the Leakey Foundation, and published in Nature, helps close one of these evolutionary gaps by revealing two early hominin species that lived side by side.

At the Ledi-Geraru site in Ethiopia’s Afar Region, an international research team discovered hominin fossils dated between 2.6 and 3.0 million years old. Lucas Delezene, an associate professor of anthropology at the University of Arkansas, served as the study’s second author, contributing alongside more than 20 scientists from North America, Africa, and Europe.

The findings include fossils of Homo that establish the oldest confirmed evidence of our lineage at 2.8 million years ago, with additional remains dating to 2.6 million years ago. These discoveries strengthen the case for Homo’s deep evolutionary roots. Even more surprising, the team found that Homo lived in the same region at the same time as another hominin, Australopithecus, around 2.6 million years ago.

This overlap challenges long-held assumptions, as Australopithecus was believed to have vanished from the area roughly 3 million years ago. The famous Australopithecus specimen known as Lucy was discovered nearby, yet her species was thought to have disappeared from the fossil record by that point.

About one in three people over the age of 80 are affected by age-related macular degeneration (AMD), and roughly 20 million Americans aged 40 and older are currently living with the condition. Most patients experience the dry form of AMD, which develops gradually and eventually leads to the loss of central vision, making it difficult to see objects directly ahead. Despite how common it is, there are still no effective treatments for the dry type of the disease.

Researchers at Aalto University have now identified a promising new method that could help stop the progression of dry AMD if applied during the early stages of diagnosis. The technique works by using heat to strengthen the cells’ natural defense systems, according to Professor Ari Koskelainen.

“Cellular functionality and protective mechanisms weaken with age, which exposes the fundus [the inside surface at the back of the eye] to intense oxidative stress,” he explains. “Free oxygen radicals damage proteins, which causes them to misfold and aggregate, then fatty protein deposits called drusen begin to accumulate, which is the main diagnostic criterion for the dry form of age-related macular degeneration.”

India's efforts to combat air pollution by using cloud seeding in its sprawling capital New Delhi appear to have fallen flat, with scientists and activists questioning the effectiveness of the move.

Cloud seeding involves spraying particles such as silver iodide and salt into clouds from aircraft to trigger rain, that can wash pollutants from the air.

Delhi authorities, working with the Indian Institute of Technology (IIT) Kanpur, began trials last week using a Cessna aircraft over parts of the city.

But officials said the first trials produced very little rainfall because of thin cloud cover.

"This will never ever do the job, it's an illusion," said Bhavreen Kandhari, an environmental campaigner in Delhi. "Only when we clean up sources of air pollution can we control it."

The government has spent around $364,000 on the trials, according to local media reports.

Each winter, thick smog chokes Delhi and its 30 million residents. Cold air traps emissions from farm fires, factories and vehicles.

Despite various interventions—such as vehicle restrictions, smog sucking towers, and mist-spraying trucks—the air quality ranks among the worst for a capital in the world.

A day after the latest trial, levels of cancer-causing PM2.5 particles hit 323, more than 20 times the daily limits set by the World Health Organization. It will likely worsen further through the season.

A study published in The Lancet Planetary Health last year estimated that 3.8 million deaths in India between 2009 and 2019 were linked to air pollution

Although the U.S. government is shut down, NASA's astronauts in space are still expected to show up for work every day.

As the shutdown continues into its second month, many federal employees are furloughed. Some, however, keep working (mostly without pay at the moment), because they're considered critical to the continued operation of the nation's functions, like the delivery of the mail.

Thankfully for the NASA astronauts aboard the International Space Station (ISS), their wellbeing falls into the category of "protection of life and safety," which NASA is tasked with maintaining while the government's doors are closed.

Like all federal agencies, NASA has had to severely cut back on its day-to-day activities, as more than 15,000 NASA civil servants have been furloughed since Oct. 1. Only essential personnel deemed "necessary to protect life and property" are granted "excepted" status, according to NASA's shutdown guidance. This includes astronauts in space and the technicians in mission control on the ground who support them.

For the most part, life aboard the ISS has continued as usual. The Expedition 73 crew currently occupying the space station have spent the past month conducting microgravity research and other experiments on their rotation and performing scheduled maintenance.