New AI Model for Drug Design Brings More Physics to Bear in Predictions

This illustration shows the mesh of anchoring points the team obtained by discretizing the manifold, an estimation of the distribution of atoms and the probable locations of electrons in the molecule. This is important because, as the authors note in the new paper, treating atoms as solid points "does not fully reflect the spatial extent that real atoms occupy in three-dimensional space."
Image Credit: Liu et al./PNAS

When machine learning is used to suggest new potential scientific insights or directions, algorithms sometimes offer solutions that are not physically sound. Take for example AlphaFold, the AI system that predicts the complex ways in which amino acid chains will fold into 3D protein structures. The system sometimes suggests "unphysical" folds—configurations that are implausible based on the laws of physics—especially when asked to predict the folds for chains that are significantly different from its training data. To limit this type of unphysical result in the realm of drug design, Anima Anandkumar, Bren Professor of Computing and Mathematical Sciences at Caltech, and her colleagues have introduced a new machine learning model called NucleusDiff, which incorporates a simple physical idea into its training, greatly improving the algorithm's performance.

X-Ray Study Reveals New Details About Betelgeuse’s Elusive Companion Star

Betelbuddy, the companion star to Betelgeuse. This image is a color composite made from exposures from the Digitized Sky Survey 2.
Image Credit: ESO/Digitized Sky Survey 2. Acknowledgment: Davide De Martin

Astronomers have long suspected that Betelgeuse — the bright red star blazing in Orion's shoulder — wasn't alone. Now, thanks to a fleeting cosmic window and swift action by Carnegie Mellon University researchers, the true nature of its elusive companion has been illuminated.

In a race against time, the CMU researchers secured director’s discretionary time on both NASA’s Chandra X-ray Observatory and the Hubble Space Telescope to investigate the long-predicted — but never detected — companion star to Betelgeuse. The timing was critical: Around Dec. 6, the companion, nicknamed “Betelbuddy,” reached its maximum separation from the massive red supergiant just before it would disappear behind it for two more years.

“It turns out that there had never been a good observation where Betelbuddy wasn't behind Betelgeuse,” said Anna O’Grady, a McWilliams Postdoctoral Fellow at Carnegie Mellon’s McWilliams Center for Cosmology and Astrophysics. “This represents the deepest X-ray observations of Betelgeuse to date.”

Retired croplands offer hope for carbon storage

An experiment at Cedar Creek Ecosystem Science Reserve tested the long-term ability of abandoned farmland to store carbon.
Photo Credit: Maowei Liang, College of Biological Sciences

Burning fossil fuels has elevated atmospheric carbon dioxide, causing massive changes in the global climate including extreme temperatures and weather events here in the Midwest. Meanwhile, human activities have increased the amount of nutrients like nitrogen and phosphorus in grasslands and forests. These are the elements in fertilizer that make lawns greener and farmlands more productive.

This overabundance of nutrients can lead to reduced water quality, the spread of invasive species and the loss of native species. However, it can also help plants capture carbon dioxide from the atmosphere and store it in the soil. This creates a paradox for environmental management: will reducing nutrient pollution make climate change worse by causing a release of carbon dioxide from the soil?

Combination of immunotherapy and targeted therapy improves survival for patients with advanced colorectal cancer

Human colorectal cancer cells
Image Credit: National Cancer Institute

A new study led by UCLA investigators found that combining zanzalintinib, a targeted therapy drug, and atezolizumab, an immune checkpoint inhibitor, helped patients with metastatic colorectal cancer, the second most common cause of cancer death in the U.S., live longer and control their disease better than with the standard treatment drug regorafenib. 

The findings simultaneously published in The Lancet and presented at the European Society for Medical Oncology Congress; mark the first time an immunotherapy-based regimen has demonstrated a survival benefit in the vast majority of patients with metastatic colorectal cancer.

“This study represents an important step forward for a group of patients who have historically had very few treatment options,” said Dr. J. Randolph Hecht, professor of clinical medicine at the David Geffen School of Medicine at UCLA and first author of the study. “We may finally be finding ways to make immunotherapy work for more patients with colorectal cancer.”

Unmasking the Culprits of Battery Failure with a Graphene Mesosponge

Photo Credit: Roberto Sorin

To successfully meet the United Nations' Sustainable Development Goals (SDGs), we need significant breakthroughs in clean and efficient energy technologies. Central to this effort is the development of next-generation energy storage systems that can contribute towards our global goal of carbon neutrality. Among many possible candidates, high-energy-density batteries have drawn particular attention, as they are expected to power future electric vehicles, grid-scale renewable energy storage, and other sustainable applications.

Lithium-oxygen (Li-O2) batteries stand out due to their exceptionally high theoretical energy density, which far exceeds that of conventional lithium-ion batteries. Despite this potential, their practical application has been limited by poor cycle life and rapid degradation. Understanding the root causes of this instability is a critical step toward realizing a sustainable and innovative energy future.

Controlling prostheses with the power of thought

 A neuroprosthesis. Artificial hands, arms, or legs can restore mobility to people with disabilities. The study investigated how the brain learns to control such prostheses via brain-computer interfaces.
Image Credit: © Sebastian Lehmann

Researchers at the German Primate Center (DPZ) – Leibniz Institute for Primate Research in Göttingen have discovered that the brain reorganizes itself extensively across several brain regions when it learns to perform movements in a virtual environment with the help of a brain-computer interface. The scientists were thus able to show how the brain adapts when controlling motor prostheses. The findings not only help to advance the development of brain-computer interfaces, but also improve our understanding of the fundamental neural processes underlying motor learning.

In order to perform precise movements, our brain's motor system must continuously recalibrate itself. If we want to shoot a basketball, this works well with a familiar basketball, but requires extra practice with a lighter or heavier ball. Our brain uses the deviations from the expected (throw) result as an error signal to learn better commands for the next throw. The brain must also perform this task when it wants to control a movement via a brain-computer interface (BCI), for example, that of a neuroprosthesis. Until now, it was unclear which regions of the brain reflect the expected result of the movement (the trajectory of the ball), which reflect the error signal, and which reflect the corrected movement command that aims to compensate for the previous error.

The Quantum Door Mystery: Electrons That Can’t Find the Exit

Photo Credit: © Technische Universität Wien

What happens when electrons leave a solid material? This seemingly simple phenomenon has eluded accurate theoretical description until now. Researchers have found the missing piece of the puzzle.

Imagine a frog sitting inside a box. The box has a large opening at a certain height. Can the frog escape? That depends on how much energy it has: if it can jump high enough, it could in principle make it out. But whether it actually succeeds is another question. The height of the jump alone isn’t enough — the frog also needs to jump through the opening.

A similar situation arises with electrons inside a solid. When given a bit of extra energy — for example, by bombarding the material with additional electrons — they may be able to escape from the material. This effect has been known for many years and is widely used in technology. But surprisingly, it has never been possible to calculate this process accurately. A collaboration between several research groups at TU Wien has now solved this mystery: just like the frog, it’s not only the energy that matters — the electron also needs to find the right “exit,” a so-called “doorway state.”

Important phenomenon discovered in the Arctic – could boost marine life

Measurements of nitrogen fixation in the Arctic Ocean aboard RV Polarstern
Photo Credit: Rebecca Duncan

Researchers from the University of Copenhagen have discovered an important phenomenon beneath the Arctic sea ice that was previously thought impossible. This phenomenon could have implications for the food chain and the carbon budget in the cold north.

The shrinking sea ice in the Arctic Ocean is, overall, a disaster. But paradoxically, the melting of the ice can also fuel the engine of the Arctic food chains: algae.

Algae are the main food source for life in the sea, but they need nitrogen to grow. And nitrogen is in short supply in the Arctic Ocean. However, a new international study led by the University of Copenhagen indicates there will probably be more of it in the future than previously thought. This could change the future prospects for marine life in the High North and possibly for the carbon budget.

Scientists Confirmed That a "Terrible" Hyena Lived in the Territory of the Modern Caucasus

The scientists used morphometric and morphological analysis of teeth.
Photo Credit: Daniyar Khantemirov

Ural scientists with colleagues from China and Azerbaijan have established that "terrible" hyenas (Dinocrocuta gigantea) lived in the territory of the modern Caucasus 10-9 million years ago. This fact was confirmed by studying jaw fragments that were found in the Upper Miocene site of Eldari, Azerbaijan. The researchers published a description and photographs of the fragments in the journal Palaeoworld.

"In our work, the Dinocrocute hyenas from the Caucasus are described for the first time. Other finds of this species are described from Southern Europe or Northern China. In other words, our finding fills a gap in understanding the distribution of dinocrocutes, which were one of the key predators in the faunas of the Miocene, a geological epoch from 23 to 5 million years before our time," explains Daniyar Khantemirov, co-author of the work, laboratory researcher at the UrFU Laboratory of Natural Science Methods in Humanitarian Research.

ADC Improves Outcomes for Patients with Advanced Triple-Negative Breast Cancer Who are Ineligible for Immune Checkpoint Inhibitors

Dr. Sara Tolaney, chief of the Division of Breast Oncology at Dana-Farber, is the senior author on the ASCENT-03 study.
Photo Credit: Courtesy of Dana-Farber Cancer Institute

Patients with an aggressive form of breast cancer who are not candidates for immune checkpoint inhibitor therapy showed significantly improved progression-free survival when treated with the antibody drug conjugate sacituzumab govitecan compared to standard chemotherapy. These findings, which stem from the ASCENT-03 trial in triple-negative breast cancer co-led by investigators at Dana-Farber Cancer Institute, are presented today at the European Society for Medical Oncology (ESMO) Congress 2025 in Berlin, Germany. They are also published simultaneously in the New England Journal of Medicine.

Triple-negative breast cancer (TNBC) accounts for about 15% of all breast cancer cases and is often difficult to treat. The 5-year survival rate for patients with metastatic disease is about 15%. Moreover, around 60% of patients with metastatic TNBC have tumors that lack the molecular marker PD-L1. This absence indicates the tumors will not respond to immune checkpoint inhibitors. For most patients with previously untreated TNBC, chemotherapy is the primary treatment option.