Knowledge editing has emerged as an attractive alternative to retraining. Conventional training requires enormous amounts of computing power, data and time, and even a small update can demand exposure to vast quantities of information. Editing methods instead attempt to identify and alter the internal parameters, representations or computational pathways associated with a particular piece of knowledge. In principle, this could allow an AI system to correct a misinformation error, learn a newly discovered scientific result or update a changing fact, such as the office held by a public figure, almost instantly. The approach promises models that can continuously adapt rather than remaining frozen at the moment their training data were collected.

The problem is that knowledge inside a large language model does not behave like a database of independent entries. A fact is typically connected to numerous other facts through relationships involving categories, causes, consequences, timelines and logical rules. If an editor changes the statement “A is located in country B,” the consequences may affect the country’s continent, neighbouring regions, languages, political institutions and thousands of related descriptions. A model that changes only the original sentence may continue producing answers based on the old information elsewhere. It could therefore state that a person holds a new position while still reasoning as if the previous office-holder remained in power, or accept a revised scientific premise while drawing conclusions from an outdated one.

This challenge becomes more serious as models move beyond simple recall. Modern systems can perform multistep deduction, connect information across documents and answer questions that require causal reasoning. They do not merely retrieve isolated sentences; they often combine multiple internal representations to generate an answer. A knowledge update that appears successful on a direct question may fail when the model is asked to reason through several related steps. The authors describe this as a need for reasoning-consistent editing: a change should not only overwrite a target fact but also preserve, revise or extend the network of inferences that depend on it.

Consider a hypothetical update stating that a medication is no longer recommended for a particular condition. A superficial edit might make the model repeat the new recommendation when directly asked. Yet the model could still claim that the medication remains standard treatment when asked about clinical guidelines, historical practice or a patient scenario. It might also produce contradictory explanations because its internal associations were only partially modified. In high-stakes fields such as medicine, law, finance and scientific research, these hidden inconsistencies could be more dangerous than an obvious lack of knowledge. A system that admits uncertainty is easier to supervise than one that confidently combines old and new information into a plausible but incoherent answer.

The first research direction proposed in the Perspective is “deductive closure circuit editing.” The idea is to identify not only the internal pathways responsible for a target fact but also the circuits that support the valid deductions connected to it. In logic, the deductive closure of a set of statements is the collection of conclusions that can be derived from those statements. Applied to language models, the concept suggests that an edit should account for the broader computational circuitry involved in generating consequences. Researchers could map how a model carries information through layers and attention patterns, then modify a coordinated group of mechanisms rather than changing a single localized representation.

Such an approach would be technically demanding because the relevant information may be distributed across many layers, attention heads and nonlinear interactions. A model may encode the same fact in several forms, including names, descriptions, patterns of association and implicit relations. Researchers would need methods to discover which pathways are genuinely causal for a model’s reasoning, rather than merely correlated with a particular output. They would also need evaluation systems that test direct recall, paraphrase, multistep inference, counterfactual reasoning and resistance to contradictory prompts. The goal would be an edit that remains stable across different wording and tasks while avoiding unintended changes to unrelated knowledge.

The second direction focuses on a model’s beliefs and confidence. Current editing procedures often treat a desired answer as unquestionably correct and attempt to force the model toward it. But models may already contain competing representations, partial evidence or uncertainty about the subject. An edit that ignores this internal landscape can erase useful information or create an artificial certainty that the evidence does not justify. Future systems could estimate how strongly a model supports different propositions, distinguish uncertainty from contradiction and use that information to determine whether an update should revise, supplement or simply contextualize existing knowledge.

Confidence, however, must be interpreted carefully. The probability assigned to a token is not the same as a calibrated belief about the truth of a proposition. A model may produce a fluent answer with high confidence because the wording is familiar, even when the underlying claim is false. Principled editing would therefore require more than reading output probabilities. It could involve comparing alternative completions, probing internal representations, checking consistency across related questions and testing whether confidence changes appropriately after new evidence is introduced. The authors’ emphasis on beliefs points toward editing methods that treat knowledge as graded and structured rather than as a collection of binary switches.

The third direction is contextualized updating, which may be essential for facts that are not universally true. Many statements depend on time, location, legal jurisdiction, scientific conditions or the source being consulted. A model should not necessarily replace an old fact with a new one when both are correct in different contexts. A law may change after a particular date; a scientific explanation may be accepted in one historical period and revised later; a person may hold an office in one year but not another. Contextualized editing would allow a model to preserve these distinctions, retrieve the appropriate version and explain why apparently conflicting statements can both be valid under different circumstances.

This shift could transform knowledge editing from a narrow repair tool into a foundation for continuously adapting AI. Instead of asking whether a model has memorized a new sentence, researchers would ask whether it has integrated an update into a coherent, traceable system of knowledge. That would require benchmarks designed around relationships, deductions and contradictions rather than simple before-and-after accuracy scores. It would also require safeguards against malicious edits, unauthorized changes and the accumulation of incompatible updates over time. If AI systems are to evolve through continual editing, they will need mechanisms for provenance, rollback, conflict resolution and independent verification.

The Perspective does not present knowledge editing as a finished technology, nor does it claim that a single method can solve the problem. Instead, it highlights a widening gap between the way current editing algorithms treat models and the way models actually use knowledge. Large language models are not conventional databases, and their internal reasoning cannot always be repaired by changing the equivalent of one record. By focusing on deductive closure, internal confidence and context-sensitive representation, the proposed research agenda offers a path toward AI systems that can learn from new information without losing logical coherence. The viral appeal of instant knowledge updates may have helped propel the field forward, but the next breakthrough will depend on making those updates dependable, explainable and consistent across everything an intelligent system can infer.

Subject of Research: Principled knowledge editing methods for reasoning-capable large language models

Article Title: Towards principled knowledge editing methods for large language model reasoning

Article References: Zhang, N., Yao, Y., Qin, J. et al. “Towards principled knowledge editing methods for large language model reasoning.” Nature Machine Intelligence (2026). https://doi.org/10.1038/s42256-026-01276-y

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s42256-026-01276-y

Keywords: Knowledge editing, large language models, AI reasoning, deductive closure, model beliefs, confidence, contextualized updates, continual learning, artificial intelligence safety

Type 1 diabetes develops when the immune system destroys the pancreatic beta cells responsible for producing insulin. Islet transplantation can restore insulin secretion by introducing healthy insulin-producing cells into a recipient, but transplanted islets are vulnerable to immune rejection and inflammatory damage. Current strategies often rely on immunosuppressive drugs, which can expose patients to infection and other complications, or on semipermeable polymer capsules intended to shield the cells from immune attack. These capsules must be thick enough to provide protection, yet thin enough to permit rapid molecular exchange. That trade-off has made long-term cell survival difficult: thicker barriers can delay the movement of glucose and insulin, while thinner barriers may offer inadequate protection or fail to form consistently around the entire cell cluster.

The new system takes its inspiration from the zona pellucida, a specialized extracellular layer that surrounds mammalian oocytes. During reproduction, this structure provides mechanical protection and participates in highly selective molecular interactions, including the recognition of sperm. Rather than reproducing the zona pellucida’s exact biological composition, the researchers adapted its design principles: a cell-associated layer is assembled through molecular recognition, formed directly at the cell surface and then strengthened into a stable hydrogel. This biomimetic concept is intended to combine the precision of biological recognition with the tunability of synthetic materials.

At the centre of the technique are aptamers, short single-stranded nucleic acid sequences that can be engineered to bind selected molecular targets. In this case, aptamer-directed recognition brings the components needed to construct the capsule into close proximity with the islet surface. Once localized, the components undergo crosslinking, a chemical process that connects individual polymer chains into a three-dimensional network. The resulting hydrogel is then hardened into a coherent coating approximately 20 micrometres thick. Because the capsule is assembled where recognition occurs, rather than being produced around the cells in a separate bulk process, it can form closely around the islet surface and avoid the uneven coverage associated with some conventional encapsulation techniques.

The researchers report a 100 per cent encapsulation efficiency, meaning that every targeted islet received a capsule under the described conditions. The encapsulation was also droplet-free. This is technically important because droplet-based microfluidic and emulsification systems can expose cells to interfaces, shear forces, changes in osmotic conditions or additional processing steps. The new procedure instead takes place in a liquid environment compatible with living cells and does not require harsh factors. According to the study, the islets experienced no detectable loss of viability or function during the encapsulation process, suggesting that the protective coating can be applied without sacrificing the biological performance it is designed to preserve.

The capsule’s thinness is another central feature. A 20-micrometre barrier is substantially thinner than many earlier protective coatings, reducing the distance that glucose must travel to reach beta cells and that insulin must cross after secretion. In principle, shorter diffusion paths can help the encapsulated islets respond more rapidly to changes in blood glucose. The hydrogel must nevertheless remain sufficiently stable to prevent direct contact between donor cells and the recipient’s immune system. The study’s design therefore treats thickness, molecular permeability and structural integrity as interconnected engineering parameters rather than as separate objectives. The researchers’ results indicate that the biomimetic coating achieved this balance well enough to support islet activity after transplantation.

The team tested the encapsulated allogeneic islets in immunocompetent diabetic mice. “Allogeneic” refers to cells transferred between genetically different individuals of the same species, a setting in which immune recognition and rejection remain relevant. “Immunocompetent” mice retain functioning immune systems, making the model more demanding than experiments conducted in animals whose immune responses have been deliberately suppressed or genetically impaired. Following transplantation, the treated animals achieved normoglycaemia, or blood-glucose levels within the normal range. Most maintained normoglycaemia for more than 100 days, indicating that the encapsulated islets continued to release biologically active insulin over an extended period despite being exposed to a functioning host immune system.

The findings suggest that the capsule did more than merely preserve the physical structure of the transplanted islets. To regulate blood glucose, beta cells must sense glucose fluctuations, produce insulin in response and release the hormone in a controlled manner. The maintenance of normoglycaemia implies that the coated islets retained sufficient glucose responsiveness and secretory capacity after transplantation. At the same time, the long duration of glycaemic control suggests that the hydrogel remained in place and continued to provide meaningful protection. The study does not establish that the coating completely eliminates immune recognition, nor does it show that the strategy would work indefinitely, but it demonstrates a sustained therapeutic effect in a stringent preclinical model.

The work also illustrates why cell protection is increasingly being approached as a problem of biological interface design. A capsule placed around a living cell is not simply a passive wall: it must regulate transport, withstand the surrounding tissue environment, avoid damaging the enclosed cells and limit harmful interactions with host biology. By using aptamers to direct assembly at the cell surface, the researchers created a process that resembles biological self-organization more closely than conventional bulk encapsulation. Such molecular precision could eventually be adapted to other cell therapies, including engineered endocrine cells or replacement cells intended to treat diseases beyond diabetes, although each application would require its own safety and compatibility testing.

For patients, the potential significance lies in the possibility of restoring insulin production without continuous systemic immunosuppression. However, the results remain preclinical. Human islets are larger, transplantation procedures are more complex and immune reactions in people can involve both cellular and antibody-mediated mechanisms. Long-term capsule stability, vascularization, inflammation, manufacturing scale-up and the fate of the material after transplantation will all need to be examined. Researchers will also need to determine whether the thin hydrogel can protect cells from the full range of immune and inflammatory threats encountered in the human body. Even so, the study presents a notable advance: a stress-free, droplet-free and molecularly guided method that creates a remarkably thin protective coating while preserving islet function. In diabetic mice with intact immune systems, that coating supported sustained glycaemic control, offering a new direction for the engineering of transplantable living cells.

Subject of Research: Biomimetic hydrogel encapsulation of pancreatic islets for immune protection and sustained blood-glucose regulation in type 1 diabetes.

Article Title: Biomimetic zona pellucida-encapsulated islets for sustained glycaemic control in immunocompetent mice.

Article References: Lee, K., Wang, X., Wen, C. et al. Biomimetic zona pellucida-encapsulated islets for sustained glycaemic control in immunocompetent mice. Nature Biomedical Engineering (2026). https://doi.org/10.1038/s41551-026-01775-8

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41551-026-01775-8

Keywords: Islet transplantation, type 1 diabetes, cell encapsulation, hydrogel, aptamers, zona pellucida, biomimetic engineering, immune protection, beta cells, glycaemic control

The study, published in Nature Human Behaviour, examined 9,367,550 weekly shopping baskets collected from more than 30,000 US households between 2017 and 2022. This long observation period allowed the researchers to compare purchasing patterns before the pandemic, during the disruption of 2020 and 2021, and in the post-pandemic period. The researchers combined household purchasing data with county-level information on pandemic policies, including restrictions that influenced mobility and retail activity. They also used artificial intelligence-assisted nutrition profiling to classify the foods purchased and assess changes in dietary quality at scale.

The results reveal a striking reversal over time. During the early pandemic and periods of lockdown, households increased the share of spending directed towards less processed foods. Such a shift may reflect the unusual circumstances of that period. With restaurants closed or operating under restrictions, families prepared more meals at home, while disruptions to daily routines may have encouraged purchases of basic ingredients and foods requiring preparation. At the same time, consumers faced uncertainty over supplies, employment and access to shops. The study indicates that the initial movement towards less processed products was not a permanent transformation of the American food basket.

In the period after the acute phase of the pandemic, spending on ultraprocessed foods rose above levels observed before COVID-19. Ultraprocessed foods are industrial formulations typically made using multiple ingredients and additives, and are often designed to be convenient, highly palatable and shelf-stable. They include many packaged snacks, ready-to-eat meals, sweetened products and processed beverages. The findings do not mean that every household adopted the same diet, nor that all processed foods have identical nutritional effects. Rather, they show that the balance of household spending moved in a direction associated with greater reliance on heavily industrially formulated products.

One of the study’s most important findings concerns how diet quality is measured. The shift was especially clear when foods were classified according to the NOVA system, which groups products by their degree of industrial processing. However, the same divergence was less pronounced when researchers used nutrient-based measures, including the Healthy Eating Index and the Food Standards Australia New Zealand nutrient profiling model. These tools assess food quality through characteristics such as nutrients, food components and overall dietary composition rather than processing alone. A product can therefore be highly processed without fitting every definition of nutritional poor quality, while a minimally processed food can still be high in salt, sugar or saturated fat.

This difference between measurement systems is scientifically important because it shows that “healthy” and “processed” are not interchangeable terms. NOVA can capture industrial transformation and the role of convenience foods in modern diets, but it does not directly measure every nutrient or dietary benefit. The Healthy Eating Index is designed to assess alignment with dietary guidance, while the FSANZ model evaluates nutritional attributes used in food profiling. By comparing all three approaches, the researchers found that each describes a different dimension of food purchasing. The study therefore supports using processing-based and nutrient-based measures together instead of treating any single score as a complete description of diet quality.

The pandemic also altered the relationship between food spending and household socioeconomic status. In 2020, the spending gap between socioeconomic groups widened, suggesting that the shock affected purchasing power and food choices unevenly. Yet the researchers report that this difference converged after the pandemic. A different pattern then appeared: a nutrition gap emerged as higher-income households increased spending on less nutritious foods. This result challenges the assumption that greater income automatically produces healthier purchasing. Resources can expand access to fresh foods and health information, but they can also increase the ability to buy convenience products, restaurant-style foods and premium ultraprocessed products.

Some food categories showed effects that persisted beyond the most disruptive months of the pandemic. Prepared foods, frozen foods and carbonated beverages were among the categories associated with lasting changes. Their persistence may reflect habits formed during periods when consumers prioritised convenience, storage life and reduced shopping frequency. Frozen and prepared products can help households manage time and uncertainty, while carbonated beverages are often heavily marketed and widely available. The continued presence of these categories in post-pandemic purchasing suggests that temporary crises can accelerate changes in the food environment that remain embedded after restrictions are lifted.

Because the analysis used purchasing baskets rather than direct records of everything people consumed, it cannot establish precisely what each person ate or prove that the pandemic alone caused every observed change. Food bought in stores may be shared, wasted, consumed outside the home or combined with ingredients purchased elsewhere. Even so, the scale and duration of the dataset provide an unusually broad view of household behaviour. By linking weekly purchases with local policies and applying several nutrition frameworks, the researchers were able to distinguish short-term disruption from longer-term change more effectively than a snapshot survey could.

The findings carry a warning for public-health planning. Food policies designed only for the emergency phase of a crisis may miss the habits that persist after the emergency ends. Support for affordable fresh foods, clear front-of-pack information, healthier prepared and frozen options, and restrictions on the promotion of products with poor nutritional profiles could help make food systems more resilient. Interventions may also need to account for socioeconomic differences without assuming that higher-income households are automatically protected from unhealthy trends. The pandemic did not produce one universal dietary response; it exposed how policy, convenience, income and food formulation interact. Understanding that complexity may be essential for preventing the next crisis from leaving behind another nutrition gap.

Subject of Research: US household food purchasing behaviours, nutrition quality and socioeconomic disparities during and after the COVID-19 pandemic

Article Title: Short- and long-term shifts in food purchasing behaviours and nutrition disparities in the USA during and after the COVID-19 pandemic

Article References:

Hu, G., Lu, W., Soberman, D. et al. Short- and long-term shifts in food purchasing behaviours and nutrition disparities in the USA during and after the COVID-19 pandemic. Nat Hum Behav (2026). https://doi.org/10.1038/s41562-026-02547-9

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41562-026-02547-9

Keywords: COVID-19 pandemic, food purchasing, ultraprocessed foods, nutrition disparities, household diets, socioeconomic inequality, NOVA classification, Healthy Eating Index, food policy, public health

Retraction is one of the strongest corrective actions used in scientific publishing. It indicates that a published paper should no longer be regarded as part of the reliable research record, even though the article may remain accessible for transparency and archival purposes. A retraction does not automatically establish that every experiment described in a paper was false, nor does it necessarily identify misconduct. It means that the article contains problems serious enough that its findings, methods, interpretation or overall reliability cannot be depended upon as originally presented. In the citation available for this case, the publication is identified as a retraction note, but no specific explanation for the decision is provided.

The withdrawn research focused on GST-NT21MP, described as a recombinant polypeptide. Recombinant molecules are produced using genetic engineering or related laboratory methods in which biological instructions are introduced into a host system so that the desired protein or peptide can be manufactured. The “GST” designation commonly refers to glutathione S-transferase, a protein frequently used in molecular biology as a fusion partner to improve the production, purification or detection of another peptide. In a fusion construct such as GST-NT21MP, the GST component may serve as a biochemical carrier or purification tag, while the NT21MP portion is intended to provide the biological activity under investigation. The retraction means that any therapeutic conclusions involving this construct must now be treated as unconfirmed.

The proposed mechanism centred on CXCR4, a seven-transmembrane chemokine receptor found on many types of cells, including subsets of breast cancer cells. CXCR4 responds primarily to the signalling molecule CXCL12, also known as stromal cell-derived factor 1. When CXCL12 binds to CXCR4, it can activate intracellular pathways involving G proteins, phosphoinositide 3-kinase, AKT, mitogen-activated protein kinases and calcium-dependent signals. In cancer biology, these pathways may support cell movement, resistance to stress, proliferation and communication with stromal cells in the tumour microenvironment. The CXCL12–CXCR4 axis has therefore attracted interest as a possible target for therapies designed to interfere with tumour dissemination or the establishment of cancer cells in distant organs.

Breast cancer is not a single disease but a collection of molecularly distinct malignancies. Tumours can differ in hormone-receptor status, HER2 expression, genomic alterations, immune-cell infiltration and sensitivity to treatment. CXCR4 activity may also vary between tumour subtypes and between cancer cells located in different regions of the same tumour. In laboratory research, an apparent reduction in CXCR4 signalling can be measured through changes in receptor abundance, downstream phosphorylation, chemotaxis, invasion through artificial membranes or tumour growth in animal models. Each assay answers a different question, and evidence from one experimental system cannot automatically establish that a molecule will be effective in patients. The withdrawal of this study removes one reported line of evidence concerning GST-NT21MP and CXCR4 in breast cancer.

For researchers, the retraction is particularly important because mechanistic claims can influence the direction of subsequent experiments. A paper proposing that a compound blocks CXCR4 may lead other laboratories to repeat the work, compare the compound with established CXCR4 inhibitors, examine its selectivity, or investigate whether the observed effects result from receptor inhibition rather than general toxicity. Reliable validation normally requires independent replication, appropriate negative and positive controls, confirmation that the fusion protein is correctly folded and biologically active, and careful separation of effects caused by the GST carrier from those caused by the therapeutic peptide. It also requires analytical evidence showing that the compound reaches the relevant cells at a biologically meaningful concentration.

The retraction also highlights the difference between a promising molecular mechanism and a clinically useful treatment. Blocking a receptor in cultured cancer cells does not demonstrate that a drug can safely reach a tumour in the human body. A candidate molecule must be evaluated for stability in blood, absorption, distribution, metabolism, elimination, immune reactions and potential toxicity. Protein-based or peptide-based agents can be rapidly degraded, may have difficulty crossing biological barriers and can trigger unintended interactions with other proteins. Even when a mechanism is valid, the therapeutic window—the difference between an effective dose and a harmful dose—must be established through progressively more rigorous preclinical and clinical studies.

The record is therefore best understood as a correction to the scientific literature rather than as evidence that CXCR4 is irrelevant to breast cancer. Research from many groups has investigated the receptor’s role in tumour biology, and the pathway remains a subject of experimental interest. However, the removal of a specific publication means that the claims attributed to GST-NT21MP should not be cited as established proof of antitumour efficacy or pathway inhibition. Scientists relying on the article should check the journal’s formal retraction record, review any accompanying editorial information and avoid using the withdrawn findings as a foundation for clinical recommendations, treatment decisions or claims about patient benefit.

For patients and the public, the central message is that a retracted research article should not be interpreted as evidence that GST-NT21MP is an available or validated breast-cancer therapy. The citation identifies a scientific publication and its correction status; it does not report a clinical trial, regulatory approval or demonstrated benefit in people. The case also illustrates how scientific publishing is designed to correct itself when concerns arise. Retractions can be unsettling, especially when a study appears to describe a novel treatment strategy, but openly marking unreliable work protects future research and helps prevent unverified findings from being amplified as medical fact. At present, the available record supports only the conclusion that the original report has been withdrawn and that its specific claims require independent reassessment.

Subject of Research: GST-NT21MP and the CXCR4 signalling pathway in breast cancer

Article Title: Retraction Note: Antitumour activity of the recombination polypeptide GST-NT21MP is mediated by inhibition of CXCR4 pathway in breast cancer

Article References: Yang, Q., Zhang, F., Ding, Y. et al. Retraction Note: Antitumour activity of the recombination polypeptide GST-NT21MP is mediated by inhibition of CXCR4 pathway in breast cancer. Br J Cancer (2026). https://doi.org/10.1038/s41416-026-03584-x

Image Credits: AI Generated

DOI: https://doi.org/10.1038/s41416-026-03584-x

Keywords: breast cancer, GST-NT21MP, CXCR4, CXCL12, recombinant polypeptide, cancer signalling, retraction, tumour biology

The target is the hydrogen 21-centimetre line, a natural radio signal produced by the spin state of neutral hydrogen atoms. In the early universe, hydrogen was the dominant element, and its atoms interacted with the surrounding radiation and matter in ways that encoded information about temperature, density and the influence of gravity. The signal is expected to contain clues about the conditions between the end of the Big Bang’s afterglow and the formation of the first stars, a period often described as the cosmic dark ages. Because the first stars had not yet begun producing light, hydrogen is one of the only available messengers from this otherwise invisible chapter of cosmic history.

Observing the signal from Earth is exceptionally difficult. As the universe expanded, the original 21-centimetre emission was stretched to much longer wavelengths, shifting it into a frequency range of approximately 10 to 50 megahertz. These frequencies are strongly affected by Earth’s ionosphere, which can absorb, distort or reflect incoming radio waves. Even when a signal reaches the ground, it is buried beneath powerful interference generated by radio stations, satellites, mobile communications and other human technologies. The Milky Way also produces intense radio emission, creating a foreground that can be millions of times brighter than the cosmological signal scientists are trying to isolate.

CosmoCube’s solution is to use the Moon as a natural radio shield. In lunar orbit, the spacecraft will spend roughly 40 minutes of every two-hour orbit above the far side, where the bulk of the Moon blocks radio noise from Earth. This temporary radio silence will not make the universe quiet, but it will remove one of the largest sources of contamination facing low-frequency astronomy. During an expected two-year mission, the satellite could collect approximately 1,000 hours of observations while shielded from terrestrial interference. By combining measurements gathered across many orbits, researchers hope to build a statistically reliable picture of the early hydrogen signal.

The spacecraft will carry a miniature, fully integrated radiometer, an instrument designed to measure extremely weak variations in radio brightness. Its antenna will unfold in lunar orbit and scan the sky for the broad spectral signature of neutral hydrogen. Rather than producing a conventional image of individual stars or galaxies, CosmoCube will measure how the average radio signal changes with frequency and direction. Those changes could reveal the thermal history of the early universe and indicate when the first sources of light began altering the surrounding hydrogen. Because the mission is expected to operate at frequencies inaccessible to most ground-based observatories, it will explore a region of cosmic history that remains largely beyond the reach of existing facilities.

Extreme sensitivity creates its own technical problems. Tiny changes in the spacecraft’s electronics, temperature or antenna response can imitate the gradual signal variations expected from the early universe. To combat this, CosmoCube will use a Dicke-switched calibration system that repeatedly compares the sky with several built-in reference sources. This process allows the instrument to track and correct internal gain changes and electronic noise. The calibration strategy is essential because the spacecraft is not simply searching for a bright, isolated transmission; it is attempting to identify a minute cosmological imprint hidden inside the instrument’s own response and the much stronger radiation of the Galaxy.

After the data return to Earth, the team will apply Bayesian statistical methods to separate the likely cosmological signal from foreground emission. Bayesian analysis allows researchers to combine measurements with physical models and calculate which possible signal histories best fit the observations. Computer simulations and in-flight measurements will also be used to reconstruct the antenna’s response to different parts of the sky. This response, known as the beam pattern, determines how much radiation from each direction contributes to a measurement. Correcting for it will help the scientists subtract distortions and distinguish genuine features of the 21-centimetre signal from artefacts introduced by the spacecraft or observing geometry.

The scientific payoff could extend beyond the history of the first stars. CosmoCube may also test how dark matter shaped the early universe. Although dark matter does not emit or absorb light in the ordinary way, its gravity helped pull hydrogen into denser regions, eventually creating the seeds of the first stars and galaxies. Different dark-matter properties would leave different signatures in the distribution and temperature of early hydrogen. By comparing the observed radio spectrum with theoretical models, scientists could investigate how rapidly matter clumped together and how the unseen component of the universe influenced the transition from darkness to Cosmic Dawn. Such measurements could complement observations from larger telescopes that study later stages of cosmic evolution.

The mission has received support from the UK Space Agency, the Kavli Foundation and the Science and Technology Facilities Council, part of UK Research and Innovation. Surrey Space Technology Limited is developing the spacecraft platform, known as SSTL-21, while researchers at the University of Cambridge, the University of Portsmouth and STFC RAL Space are contributing to the science and engineering. Partners from European countries, including Malta, are also involved. Working laboratory prototypes have already been built, and representative spacecraft and payload models are undergoing environmental testing to assess whether the instrument can maintain the required thermal and electronic stability in lunar orbit. The team has participated in the European Space Agency’s mini-Fast mission call, with a proposed mission cost below €50 million, and hopes CosmoCube could launch within the next five years.

CosmoCube will not be the only project seeking shelter behind the Moon. Space agencies in the United States, India and other countries are developing concepts for lunar-orbiting or lunar-surface radio observatories, recognizing that the far side offers a rare zone protected from Earth’s technological noise. What makes the British proposal especially notable is its compact scale: a relatively small spacecraft is being designed to perform precision cosmology in one of the most demanding environments accessible to modern space science. If its calibration systems and analysis techniques work as planned, CosmoCube could turn the Moon into a listening post for the universe’s earliest audible trace, bringing scientists closer to understanding how the first stars emerged from the darkness.

Subject of Research:
The cosmic dark ages, the 21-centimetre hydrogen signal, Cosmic Dawn and the influence of dark matter on the formation of the first stars and galaxies.

Article Title:
The CosmoCube Lunar Mission for Probing the Dark Ages and Cosmic Dawn via 21-cm Cosmology

News Publication Date:
14-Aug-2026

Web References:
https://www.nature.com/articles/s41550-026-02946-y

References:
Nature Astronomy, DOI: 10.1038/s41550-026-02946-y

Image Credits:
Surrey Space Technology Ltd

Keywords

CosmoCube, lunar mission, cosmic dark ages, Cosmic Dawn, 21-centimetre cosmology, neutral hydrogen, dark matter, radio astronomy, Moon, University of Cambridge, low-frequency astronomy, early universe

Lead poisoning most often occurs when people swallow contaminated water, food, household dust, or particles of deteriorating paint. Children are particularly vulnerable because their developing nervous systems can be permanently damaged by even relatively small amounts of exposure. Lead can interfere with brain development, learning, behavior, and other physiological processes, and it may enter drinking water when old plumbing systems release the metal into the supply. Current interventions focus primarily on identifying and eliminating the source of contamination. For patients with severe poisoning, doctors may use chelation therapy, in which chemical agents bind lead and help the body excrete it. Such treatment, however, is not generally intended as a routine defense against low-level environmental exposure.

Farny’s team is investigating whether a probiotic bacterium could perform a similar binding function within the gastrointestinal tract. The researchers will concentrate on Escherichia coli Nissle, a nonpathogenic strain used in probiotic products outside the United States to help treat gastrointestinal conditions, including diarrhea. Unlike disease-causing strains of E. coli, Nissle has a long history of use in certain medical and nutritional applications. In the WPI study, the bacterium will be modified to produce aptamers, short strands of DNA or RNA capable of folding into structures that recognize and attach to specific molecules or ions. The central idea is to make the bacteria act as microscopic lead traps while they pass through the digestive system.

The first experiments will be conducted in Caenorhabditis elegans, a soil-dwelling nematode that measures approximately one millimeter in length and feeds on bacteria. C. elegans is widely used in biology because its transparent body, short life cycle, and well-characterized genetics allow scientists to observe biological processes efficiently. Though far simpler than a human, the worm possesses conserved cellular and molecular pathways that can provide an initial indication of whether a biological intervention is toxic, effective, or capable of changing the movement of substances through the gut. Farny’s group will feed the worms engineered bacteria together with lead and then examine whether the metal remains bound within the intestinal contents rather than entering the animals’ tissues.

The proposed mechanism depends on keeping lead chemically sequestered long enough for it to leave the body. Aptamers can be designed or selected to fold into three-dimensional shapes with binding sites that attract particular targets. If the engineered E. coli Nissle produces these molecules in the worm’s gastrointestinal tract, the aptamers could attach to lead ions and reduce their availability for absorption. The researchers will need to determine whether the molecules remain stable in the gut, whether they bind lead strongly in the presence of other minerals and digestive compounds, and whether the bacteria can produce enough aptamer to make a measurable difference. They must also verify that the engineered microbes do not create unintended biological effects.

A second part of the project will use artificial intelligence to search for a potentially simpler route to lead binding. Dmitry Korkin, WPI’s Harold L. Jurist ’61 and Heather E. Jurist Dean’s Professor of Computer Science, will analyze genetic information from E. coli Nissle to identify sequences that might already encode useful metal-binding capabilities. Rather than inserting entirely new genetic material, the researchers hope to rearrange or repurpose existing bacterial code. This approach could reduce the complexity of engineering the organism and potentially simplify later development, although any candidate design would still require extensive laboratory testing for performance and safety.

Korkin describes the process as an iterative discovery loop in which computational predictions guide experiments and experimental results improve subsequent predictions. Machine-learning systems can compare large collections of genetic sequences, identify patterns associated with biomolecule production, and suggest combinations that might generate efficient binding structures. In this project, AI will not replace laboratory validation. Predicted genetic arrangements must be introduced into bacteria, tested under controlled conditions, and evaluated for how strongly and selectively their products bind lead. Results from those experiments can then be fed back into the computational models, allowing researchers to refine the designs over multiple cycles.

The project builds on years of work in Farny’s laboratory, including student-led research supported by early funding from Robert F. Ferrari, a WPI alumnus and president of Northeast Water Solutions, a Rhode Island company specializing in water systems engineering. Farny is a synthetic biologist whose research examines how biological systems can be adapted to address environmental challenges. Her previous work includes research into gene regulation in soil bacteria used in industrial and environmental engineering, supported by a 2024 National Science Foundation CAREER Award. She is also a co-inventor on four patent applications involving aptamers, giving the new project a foundation in both microbial engineering and nucleic-acid-based molecular design.

The researchers emphasize that the work is not yet a treatment for lead poisoning and that results from C. elegans cannot be directly translated into human use. A successful experiment in worms would represent only an early proof of concept. Any future probiotic designed to bind lead would need to undergo studies in more complex animal models, followed by rigorous evaluation of dosage, stability, efficacy, microbial safety, genetic containment, and interactions with the human microbiome. Scientists would also need to establish that captured lead is reliably eliminated rather than merely redistributed within the body. Nevertheless, a living bacterial system could eventually complement environmental remediation by offering temporary protection while contaminated pipes, soil, paint, or water sources are being addressed. The NIH-supported study, funded through the National Institute of Environmental Health Sciences under award R21ES038018, is therefore testing whether engineered microbes can become a new biological tool against one of the world’s most persistent environmental health threats.

Subject of Research: Engineered probiotic bacteria and aptamers for reducing gastrointestinal lead absorption

Article Title: Engineered Probiotic Bacteria Could Help Trap Lead in the Gut

Web References: https://www.wpi.edu/people/faculty/nfarny; https://www.wpi.edu/academics/departments/biology-biotechnology; https://web.archive.org/web/20230129015832/https:/www.wpi.edu/people/faculty/dkorkin; https://www.sciencedirect.com/science/article/pii/S1871678424000256; https://wp.wpi.edu/farnylab/; https://www.wpi.edu/news/wpi-researcher-receives-12-million-award-determine-how-environment-impacts-gene-expression-bacteria

References: National Institute of Environmental Health Sciences, National Institutes of Health, award R21ES038018; Worcester Polytechnic Institute research announcement

Image Credits: Worcester Polytechnic Institute

Keywords

Lead poisoning, engineered probiotics, E. coli Nissle, aptamers, C. elegans, synthetic biology, bioengineering, artificial intelligence, environmental health, public health, gastrointestinal tract, microbial engineering

The findings, published in Cell, suggest that polyamines act as molecular storage lockers for iron, binding or buffering the metal in a form that is less likely to trigger destructive chemistry. The work helps explain a long-standing biological puzzle. Cells maintain large pools of polyamines even though their best-known functions, including helping RNA fold and supporting cell growth, appear to require only a fraction of the total amount present. According to the researchers, these molecules serve an additional and previously underappreciated purpose: they help control the pool of chemically reactive iron that can otherwise push cells toward ferroptosis, a form of iron-dependent cell death caused by the breakdown of membrane lipids.

The research began with a broader question about why polyamines are so abundant. The Jain Lab studies RNA, the molecule that carries genetic instructions from DNA to the protein-making machinery and performs important structural and regulatory roles in its own right. Polyamines bind to RNA and can influence how it folds, but Ankur Jain, Whitney Henry, Pushkal Sharma, and their colleagues suspected that this explanation could not account for the extraordinary quantities of polyamines maintained by living cells. Their concentration can approach that of ATP, the molecule that supplies energy for many cellular reactions. Cells deprived of polyamines stop growing and dividing, but the known RNA-related functions do not fully explain why losing them is so damaging.

To search for hidden functions, the researchers used a genome-wide genetic screen. Rather than examining individual genes one at a time, this approach allows scientists to perturb or assess thousands of genes across a cell population and identify which pathways become especially important under particular conditions. The screen revealed that cells with reduced polyamine levels became unusually dependent on GPX4, an enzyme that protects cell membranes from oxidative damage. GPX4 converts toxic lipid hydroperoxides into less harmful compounds, preventing the chain reactions that can cause membranes to rupture. When this protection fails in the presence of reactive iron, cells can undergo ferroptosis.

The genetic results pointed toward a connection between polyamine loss, iron chemistry, and membrane damage. Cells with fewer polyamines also accumulated larger amounts of a protein that functions as an iron sponge, storing the metal in a mineralized and comparatively inert form. The response suggested that the cells were compensating for the loss of another iron-buffering system. The researchers proposed that polyamines themselves might normally keep iron in a safe state, reducing the amount available to catalyze harmful reactions. This distinction is important because total cellular iron is not necessarily a measure of iron toxicity. The dangerous fraction is labile or chemically reactive iron, which can move between molecular partners and participate in oxidation reactions.

To test the hypothesis directly, the team developed a fluorescent sensor capable of detecting chemically reactive iron inside living cells. The sensor changes the way cells glow according to the amount of reactive iron present, allowing researchers to follow iron chemistry in real time under a microscope rather than relying only on measurements from disrupted cells. The scientists combined this tool with an earlier sensor that reports polyamine levels. Watching both signals simultaneously revealed a striking inverse relationship. As polyamine concentrations fell, the amount of chemically reactive iron rose. The result provided direct evidence that polyamines help suppress the buildup of the iron fraction most capable of damaging cellular components.

The discovery also clarifies how iron, polyamines, and ferroptosis may be connected at the molecular level. When reactive iron is available, it can accelerate the oxidation of polyunsaturated fatty acids embedded in cell membranes. These oxidized lipids can initiate a self-amplifying chain reaction, progressively weakening the membrane. GPX4 interrupts that process by removing lipid peroxides, but cells with diminished polyamines appear to place greater reliance on the enzyme because they contain more reactive iron. In this model, polyamines act earlier in the sequence by limiting the chemical availability of iron, while GPX4 provides a second protective barrier against the lipid damage that follows when iron-driven oxidation occurs.

The findings could be especially relevant to cancer biology. Many cancer cells maintain elevated polyamine levels to support rapid growth, protein production, and cell division. Drugs that interfere with polyamine synthesis or transport have therefore been investigated as potential cancer treatments, but reducing polyamines alone has often produced limited results. The new work suggests one reason: when polyamines decline, cancer cells may activate compensatory iron-storage pathways and become dependent on GPX4 for survival. Combining a treatment that lowers polyamine levels with a drug that blocks GPX4 could, in principle, remove two protective systems at once, allowing reactive iron and lipid peroxidation to reach lethal levels. The researchers emphasize that this therapeutic strategy will require careful testing, since normal cells also depend on both iron regulation and antioxidant protection.

The biological implications may extend beyond cancer. Mutations affecting the movement and regulation of polyamines have been linked to a rare form of early-onset Parkinson’s disease, while abnormal iron accumulation has long been observed in the brains of people with Parkinson’s disease. The new findings do not establish that reactive iron caused by disrupted polyamine metabolism is responsible for neuron loss, but they provide a possible mechanistic connection between the two observations. Neurons are particularly vulnerable to oxidative damage because they are metabolically active and contain membranes rich in oxidation-sensitive lipids. If polyamine distribution or production is impaired, the resulting increase in labile iron could place additional pressure on antioxidant systems and contribute to cellular stress.

Beyond the proposed disease links, the fluorescent iron sensor may become one of the study’s most widely useful products. Researchers investigating aging, metabolism, cancer, neurodegeneration, and cellular stress can use it to observe reactive iron in living cells and compare its behavior with other molecular signals. The ability to measure labile iron dynamically could reveal when and where the metal becomes dangerous, rather than simply showing that total iron levels have changed. For now, the study recasts polyamines as more than helpers of RNA structure and cell proliferation. They appear to be part of a fundamental chemical defense system, quietly buffering one of biology’s most essential—and potentially destructive—metals.

Subject of Research: Polyamines, reactive iron, ferroptosis, cellular iron buffering, cancer biology, and neurodegeneration

Article Title: Polyamines buffer labile iron to suppress ferroptosis

Web References: https://mediasvc.eurekalert.org/Api/v1/Multimedia/ea90bd0f-d966-462c-a88a-12240ea8ea37/Rendition/low-res/Content/Public

References: Cell, “Polyamines buffer labile iron to suppress ferroptosis,” publication date: 14-Aug-2026

Image Credits: Pushkal Sharma / Whitehead Institute

Keywords: polyamines, labile iron, ferroptosis, GPX4, iron toxicity, cancer treatment, Parkinson’s disease, reactive oxygen species, lipid peroxidation, Whitehead Institute

The research, published in the Journal of Geophysical Research: Machine Learning and Computation, focuses on active regions: magnetically complex areas of the Sun where sunspots form and powerful flares or coronal mass ejections may eventually originate. These regions do not appear instantly. Magnetic fields generated deep inside the Sun gradually rise through its turbulent interior and begin interacting with the visible surface over a period of several hours. Their complete development can take one or more days. During the earliest stage, the magnetic structure remains hidden from direct observation, making the emergence difficult to identify with conventional monitoring techniques. EarlyDetect attempts to solve that problem by learning how hidden magnetic evolution changes measurable signals at the surface.

The model was developed by NJIT undergraduate researcher Jonas Tirona with solar physicists and computer scientists at NJIT, Princeton University and NASA’s Ames Research Center. The team trained it using observations from NASA’s Solar Dynamics Observatory, or SDO, a spacecraft that continuously monitors the Sun. In particular, the researchers used data from SDO’s Helioseismic and Magnetic Imager, which records the Sun’s surface magnetic field and Doppler velocity every 45 seconds. Doppler measurements reveal motions toward and away from the spacecraft, allowing scientists to study the acoustic waves that travel through the solar interior. By processing these observations into hourly maps, the team created a time-dependent view of how a potential active region changes before it becomes visible.

The model does not simply look for a new dark spot. Instead, it analyzes several related signals and searches for patterns that precede emergence. One of the earliest signatures appears in acoustic power, a measurement of the strength of solar oscillations inferred from Doppler-velocity observations. As magnetic fields rise through the Sun, they can alter the propagation and behavior of acoustic waves. These changes may occur before the magnetic concentration is strong enough to produce a conspicuous surface feature. A later sequence of changes can appear in continuum intensity, which measures the Sun’s visible brightness, followed by more direct alterations in the magnetic field. EarlyDetect is designed to connect these weak, evolving signals rather than treating each observation as an isolated event.

To accomplish this, the researchers used a Transformer architecture, a class of machine-learning model best known for powering large language systems such as ChatGPT. Transformers are effective at analyzing sequences because they can compare information across widely separated points in time and determine which changes are most important. In EarlyDetect, the sequence is not made of words but of solar observations arranged chronologically. The model learns relationships among acoustic power maps, continuum-intensity data and magnetic-field measurements, then estimates whether the observed pattern is likely to develop into an emerging active region. This approach allows the system to consider both the strength of a signal and the order in which different signals appear.

One of the study’s most unexpected findings came from an attempt to make the data easier for the model to interpret. The researchers applied a filtering technique intended to suppress noise and emphasize broader trends. Instead, the filtering reduced the model’s performance. By averaging away rapid, low-amplitude fluctuations, it removed some of the faint variations that carried the earliest information about an emerging region. Alexander Kosovichev, distinguished professor of physics at NJIT and co-principal investigator, compared the problem to trying to detect a slight change in rhythm within a noisy orchestra. In this case, what appeared to be noise may have contained precisely the subtle short-timescale variations that distinguished an emerging region from ordinary solar turbulence.

After training on SDO/HMI observations, the researchers evaluated EarlyDetect using active regions that had not been included in the model’s training data. The best-performing version identified precursor signals an average of 9.24 hours before the regions became visible. It outperformed both a standard Transformer model and an earlier benchmark method used for comparison. The result suggests that advanced machine learning can extract predictive information from solar measurements that may be too faint, complex or variable for traditional analysis methods to identify consistently. However, the researchers emphasize that the reported average is not a guaranteed warning time for every event. Some regions may be detected earlier, while others may generate a late warning or no reliable warning at all.

The potential importance of earlier detection extends beyond predicting when a sunspot will appear. Active regions can produce solar flares, which release intense bursts of radiation, and coronal mass ejections, which propel clouds of magnetized plasma into space. When directed toward Earth, these events can disturb the planet’s magnetosphere, degrade high-frequency radio communication, interfere with satellite operations and induce electrical currents in long transmission lines. An alert that a magnetically active region is developing could give satellite operators, communication providers and power-grid managers additional time to review protective procedures. Yet the emergence of an active region is not itself a forecast that a major flare or coronal mass ejection will occur. Many active regions remain relatively quiet, and the relationship between emergence and later eruptions is not simple.

The researchers therefore describe EarlyDetect as a promising research prototype rather than an operational space-weather forecasting system. It was trained on known emergence events and still produces false alarms and late predictions. Its performance must be tested across a much larger and more diverse collection of solar events, including different phases of the Sun’s approximately 11-year activity cycle. The team has released the Solar Active Region Emergence Dataset, known as SolARED, to support that work. The dataset compiles observations from SDO and is accompanied by the Solar Active Region Portal, an interactive platform that allows researchers to explore the observations and develop alternative prediction methods. By making the data publicly available, the team hopes to create a shared benchmark for both heliophysics and artificial-intelligence research.

The project was among the first research efforts supported by NJIT’s Grace Hopper AI Research Institute, launched in 2025 to encourage interdisciplinary work in artificial intelligence. Additional support came from NASA heliophysics and space-weather research programs, including the NASA Science DRIVE Center Consequences Of Fields and Flows in the Interior and Exterior of the Sun. For Tirona and his collaborators, the study demonstrates how machine learning can turn continuous streams of solar measurements into early clues about events hidden beneath the Sun’s surface. A dependable operational system remains a distant goal, but the ability to identify active-region emergence hours before it becomes visible offers a new way to investigate the Sun—and could eventually provide an earlier line of defense against its most disruptive outbursts.

Subject of Research: Solar active-region emergence and machine-learning-based space-weather forecasting

Article Title: Forecasting Continuum Intensity for Solar Active Region Emergence Prediction Using Transformers

News Publication Date: 14-Aug-2026

Web References: https://doi.org/10.1029/2025JH001207; https://sun.njit.edu/sarportal/

References: Journal of Geophysical Research: Machine Learning and Computation, DOI: 10.1029/2025JH001207

Image Credits: Irina N. Kitiashvili, NASA Ames Research Center, and Spiridon Kasapis, Princeton University

Keywords

Artificial intelligence, machine learning, solar physics, solar active regions, space weather, solar magnetic fields, helioseismology, Transformer models, NASA Solar Dynamics Observatory, solar flares, coronal mass ejections, Sun

The research was conducted through the EUROlinkCAT collaboration, an international programme examining the health and development of children born with congenital anomalies. Congenital anomalies, sometimes called birth defects, are structural or functional abnormalities that develop during pregnancy and are present at birth. They include conditions affecting the heart, brain, spine, limbs, digestive system and other organs. Although advances in surgery, neonatal medicine and long-term care have substantially improved survival, many affected children require repeated medical appointments, specialist treatment, rehabilitation and support throughout childhood.

To investigate whether social circumstances influenced survival, the researchers securely linked congenital anomaly registry information with routinely collected health and administrative records. These records can include hospital admissions, primary-care contacts, mortality data and indicators of socio-economic status. The analysis covered participating regions in Wales, England, Northern Ireland, Italy, Finland, Ukraine, Malta, Spain and Denmark. Because the data were linked within each country under strict privacy and governance arrangements, researchers could study long-term outcomes without exposing individuals’ identities. The approach demonstrates how population-scale data can reveal patterns that would be difficult to detect through individual clinical studies.

The results showed a clear social gradient in mortality. During the first year of life, children with major congenital anomalies from the most socio-economically disadvantaged groups had a 47 per cent higher risk of death than children from the most affluent groups. The disparity became more pronounced between the ages of one and ten, when mortality was approximately twice as high among the most disadvantaged children. In statistical terms, the findings indicate that deprivation was associated with a substantially elevated hazard of death, even though the size and consistency of the association differed between countries and regions.

The researchers suggest that the changing pattern across childhood may provide important clues about how inequality affects health. During infancy, children with severe anomalies often receive intensive and highly specialised treatment in neonatal units, paediatric hospitals and surgical centres. Such services may operate according to national or regional protocols, potentially reducing differences in access during the earliest and most medically urgent period. After the first year, however, care often becomes more dependent on general practitioners, community nurses, therapists, social services and families’ ability to attend appointments and coordinate complex treatment. Housing conditions, transport, household income, health literacy and access to nutritious food may therefore have a stronger influence on outcomes.

Professor Sue Jordan, the study’s lead author, said differences in mortality between rich and poor were greater after the first year of life, when care depends less on acute services and more on primary and community care. She noted that the inequalities were consistently observed across comparisons in Wales and Ukraine, the countries in the study with the lowest gross domestic product per capita. The authors hope the findings will prompt closer examination of the resources available to children with congenital anomalies who are living in poverty, particularly after they leave intensive early-life medical services.

Wales was among the participating regions where socio-economic disadvantage was consistently associated with higher mortality during infancy and later childhood. The Welsh analysis was enabled by the SAIL Databank, based at Swansea University and funded by the Welsh Government. SAIL provides approved researchers with secure access to anonymised, linkable population-scale data. Its architecture allows information from different public-sector sources to be analysed together while maintaining controls over disclosure, access and data governance. For congenital anomaly research, this creates an opportunity to follow children beyond the initial diagnosis and examine outcomes across multiple stages of development.

The international comparison also produced a potentially important exception. The association between deprivation and mortality was not observed in Denmark, suggesting that the inequalities identified in other settings are not inevitable consequences of having a congenital anomaly. The researchers caution that differences in health systems, social protection, data coverage, anomaly profiles and the measurement of deprivation may all contribute to variation between countries. Nevertheless, Denmark’s result raises the possibility that policies reducing the effects of poverty on children’s health, combined with coordinated medical and community care, could offer lessons for other health systems.

The study identified additional social patterns. In countries where nationality information was available, children born to non-European Union nationals experienced higher mortality. The analysis did not find a clear association between maternal marital status and childhood mortality. These findings do not establish that nationality itself causes poorer outcomes; instead, nationality may reflect wider factors such as migration status, language barriers, insecure housing, limited knowledge of health services, financial hardship or difficulties accessing care. The authors emphasise that secure linkage of routinely collected records can help identify such patterns and guide more targeted research into the mechanisms behind them.

Dr Hywel Turner Evans of the SAIL Databank said the findings showed that children with major congenital anomalies faced a higher risk of death when they came from more disadvantaged backgrounds, while the absence of the same pattern in Denmark suggested that these inequalities could be reduced. The researchers argue that improving survival will require more than advances in surgery and hospital medicine. Earlier identification of families needing additional support, better coordination between specialist and community services, and sustained investment in public health and social care may be essential. By bringing together data from several European health systems, the EUROlinkCAT study provides evidence that a child’s social environment can shape survival long after the immediate medical crisis of birth has passed.

Subject of Research: People

Article Title: Mortality amongst European children with congenital anomalies: Associations with socio-economic status in the EUROLINKCAT cohort

Web References: https://doi.org/10.1371/journal.pone.0352025

References: PLOS One, DOI: 10.1371/journal.pone.0352025

Keywords: congenital anomalies, childhood mortality, socio-economic inequality, poverty, EUROlinkCAT, SAIL Databank, Wales, child health, public health, European health data

The research, led by engineer Kaifeng Zhang of the Foshan Meteorological Bureau under the guidance of Professor Lingkun Ran of the Institute of Atmospheric Physics at the Chinese Academy of Sciences, reconstructs a tornado that struck Guangdong Province in June 2022. The team used a Weather Research and Forecasting, or WRF, model configured at an exceptionally fine horizontal resolution of 37 meters. At this scale, the model can represent much smaller features of a convective storm than conventional regional forecasting systems, including concentrated zones of rotation, sharp wind gradients, and localized updrafts. The findings, published in Atmospheric and Oceanic Science Letters, show that incorporating observations from a nearby X-band phased-array radar substantially improved the simulated tornado.

The radar used in the study was the Foshan Nanhai X-band dual-polarization phased-array radar. X-band systems operate at relatively short wavelengths, allowing them to detect detailed structures in precipitation and storm-scale wind fields, particularly when positioned close to the event. A phased-array radar can also scan the atmosphere rapidly by steering its beam electronically rather than relying entirely on mechanical rotation. This rapid sampling is valuable for tornado research because the circulation can intensify, shift, or weaken during the short interval between conventional radar scans. Dual-polarization measurements provide additional information about the size, shape, and composition of particles within the storm, helping researchers distinguish rain, hail, and other hydrometeors.

To bring these observations into the atmospheric model, the researchers applied a three-dimensional variational data assimilation technique, commonly known as 3D-Var. Data assimilation combines a model’s preliminary estimate of the atmosphere with real-world observations, producing an updated analysis that is intended to be closer to the actual state of the storm. In this case, the radar supplied two important forms of information: radial velocity, which measures the component of wind moving toward or away from the radar, and reflectivity, which indicates the strength of the returned radar signal from particles inside the storm. The assimilation process adjusted the model’s wind and moisture-related variables so that the simulated storm more closely matched the observed radar structure.

The team compared four experiments to determine which types of radar information contributed most strongly to the simulation. The control experiment, known as CTRL, used no radar data. The REF experiment assimilated radar reflectivity, while the VEL experiment used radial wind observations. The XPAR experiment incorporated both radial velocity and reflectivity from the phased-array radar. These controlled comparisons allowed the researchers to separate the effects of dynamical information, represented primarily by wind measurements, from microphysical information associated with precipitation and hydrometeor structure.

The results revealed a clear difference between the experiments. When radial wind data were included, the simulated storm developed a stronger and more coherent rotational circulation. This improvement is important because a tornado is not simply a column of intense rain; it is a concentrated dynamical vortex whose wind field must be correctly organized across several levels of the atmosphere. Radar radial velocity does not measure the complete three-dimensional wind directly, but it provides critical evidence of converging and rotating air along the radar beam. Assimilating those observations helped the model establish a more realistic low-level vortex, strengthening the spinning motion associated with the observed tornado.

Reflectivity data contributed in a different but complementary way. Reflectivity describes how strongly precipitation particles scatter the radar signal, offering clues about the distribution and intensity of rain, hail, and other particles within the storm. Although reflectivity alone does not directly reveal the wind circulation, it helps define the storm’s microphysical environment, including where precipitation is forming and how it is being transported. In the study, reflectivity assimilation improved the representation of the storm’s precipitation structure, but it was less effective than radial velocity at creating the low-level dynamical circulation required for a realistic tornado. The findings indicate that precipitation structure and wind structure must be treated as related but distinct elements of the simulation problem.

The experiment that combined both observations, XPAR, produced the most accurate representation of the tornado’s track. Its predicted path closely matched the observed trajectory, outperforming the VEL, REF, and CTRL experiments. The ranking was XPAR first, followed by VEL, REF, and CTRL. This ordering suggests that radial wind data were the dominant ingredient for capturing the tornado’s motion and rotation, while reflectivity added valuable information that refined the surrounding storm environment. Together, the two data types gave the model a more complete picture: radial velocity helped define the storm’s dynamics, and reflectivity helped constrain the distribution of precipitation and the associated microphysical processes.

The study is significant because weak tornadoes in westerly environments can be especially difficult to identify and simulate. Mid-latitude westerly flow can transport weather systems rapidly, while local variations in wind direction and speed with height may influence the development of rotation. A weak tornado may also produce faint or short-lived radar signatures compared with a large, classic supercell tornado. By using an ultra-high-resolution model and rapidly updated local radar observations, the researchers were able to examine how a small-scale vortex formed within a broader weather system. Professor Ran said the experiment quantifies the value of phased-array radar for mesoscale meteorology and demonstrates that radar wind assimilation is essential for building a realistic low-level vortex, while reflectivity data help refine the storm’s microphysical environment.

The researchers emphasize that the work represents a case study rather than a complete forecasting solution. A single tornado cannot establish whether the same assimilation strategy will perform equally well across different storm types, terrain conditions, radar distances, and atmospheric environments. The team plans to test the method on a wider range of tornado events to assess its reliability for regional numerical weather prediction. If the results can be generalized, rapid-scan phased-array radar combined with high-resolution data assimilation could improve short-term forecasts and warning decisions, giving emergency managers and the public more precise information about where a tornado may travel. For communities in the Pearl River Delta and other densely populated regions, even modest improvements in the timing and location of warnings could reduce exposure to one of nature’s fastest and most destructive hazards.

Subject of Research: Tornado simulation and phased-array radar data assimilation

Article Title: X-band phased-array radar data assimilation for a case study on a weak tornado numerical simulation in the westerlies

News Publication Date: 22 May 2026

Web References: https://doi.org/10.1016/j.aosl.2026.100863

References: Atmospheric and Oceanic Science Letters, DOI: 10.1016/j.aosl.2026.100863

Image Credits: Kaifeng Zhang

Keywords: Tornadoes, extreme weather, weather simulations, X-band radar, phased-array radar, radar data assimilation, numerical weather prediction, atmospheric science, mesoscale meteorology, Guangdong, Pearl River Delta