Growing food more sustainably: what’s the best way to do this? It is one of the big challenges: producing enough food without compromising the soil. After all, healthy soil has many functions — called multifunctionality — and for sustainable agriculture these must be preserved.

‘A multifunctional soil is essential for sustainable food production, because plants get their food from it,’ state the researchers from NIOO and Wageningen University & Research (the Netherlands), and the Universität Tübingen (Germany). ‘Soil also has indispensable roles in water storage, coping with climate change and disease suppression.’

Organic vs conventional

Research on farms across the Netherlands now shows that it is mainly the intensity of tillage that determines whether the soil can retain all its functions. Interestingly, the difference between conventional and organic farming has less of an influence. In both types of agricultural systems, a lot of variation is found in soil tillage and management.

‘The good news is that in conventional agriculture, which is the vast majority, there is a lot to gain,’ states soil ecologist and NIOO professor Wim van der Putten. ‘On all farms, including organic ones, it is important at this point not to cultivate the soil too intensively. For example: ploughing less. Inverting the soil during ploughing is a very big disruption for soil life.’

More than ploughing less

Not only less frequent ploughing but also making more use of mixtures of grasses and plants from the bean family, such as clovers, contributes to multifunctional healthy soil. You can alternate these with growing cereals such as wheat, barley, spelt and rye.

The research team took samples and carried out measurements at more than 50 Dutch agricultural farms on both clay and sandy soils. This was always done in pairs: a farm with conventional agriculture plus an organic neighbouring farm. The soil type and other conditions were then very similar. ‘That way, we could compare them like twins,’ clarifies Guusje Koorneef. Together with Sophie van Rijssel, she conducted her PhD research on this topic.

Sustainable and productive

A wide array of soil properties was measured and farmers shared what farming practices they applied. The organic carbon present in the soil proved to be the best predictor of soil multifunctionality, and for biological indicators this was the bacterial biomass. Koorneef adds: ‘We looked at both sandy and marine clay soils. These are two very different soil types in the Netherlands. And we see the same picture in both soil types.’

‘The popular term of sustainable intensification is contradictory to our results,’ argues contributing researcher Kyle Mason-Jones, now working at the Universität Tübingen. ‘More intensive soil management leads to reduction of soil functions and is thus less sustainable.’ Therefore, the researchers propose a new, appropriate goal. ‘Productive de-intensification. If it is successful, you will get more functions from a less intensively cultivated soil while retaining the crop yield as much as possible.’

Vital soil

These findings are the final result of the Vital Soils project. The project was subsidised by NWO Groen, coordinated by NIOO and carried out together with Wageningen University & Research. Besides the scientific partners, there were also several social partners involved: Eurofins-Agro, BO Akkerbouw, Open Teelten (formerly PPO-AGV) and LTO-Noord.

Previous research using satellite imagery, within the same project, measured the ‘greenness’ of crops in the field. This gives us an estimate of production levels. It showed that the degree of greenness (the crop yield) did not suffer from a decrease in management intensity. Interestingly, organic farming could return to being as productive as conventional farming about 17 years after the transition.

Back to the current research. ‘You don’t necessarily have to have gone through the entire transition to organic farming to still have a positive impact on soil health,’ says Koorneef. ‘I find it really promising that in both conventional and organic farms you can strengthen the functioning of the soil by working it less intensively.’

New findings may explain why: Not all senescent skin cells are the same.

Researchers from Johns Hopkins University have identified three subtypes of senescent skin cells with distinct shapes, biomarkers, and functions — an advance that could equip scientists with the ability to target and kill the harmful types while leaving the helpful ones intact.

The findings were published today in the journal Science Advances.

“We’ve known that senescent skin cells are different from senescent immune cells or senescent muscle cells. But within a cell type, senescent cells are often considered the same — in essence, skin cells are either senescent or not, for example,” said Jude Phillip, an assistant professor of biomedical engineering at Johns Hopkins University. “But we’re finding that when a skin cell goes into senescence, or a zombie-like state, the cell could go down one of three different paths, each leading to a slightly different subtype.”

Leveraging new advances in machine learning and imaging technology, the researchers compared skin cell samples from 50 healthy donors between the ages of 20 and 90 who participated in the Baltimore Longitudinal Study, an NIH-funded project that is the longest ongoing study of aging in the United States.

Researchers extracted fibroblasts — cells that produce the scaffolding to give tissues their structure — associated with skin tissue and pushed them toward senescence by damaging their DNA, something that happens with aging. Because senescent cells build up naturally as people grow older, the aged samples contained a mix of healthy/nonsenescent and senescent fibroblasts.

Using specialized dyes, the researchers were able to capture images of the cells’ shapes and stained elements that are known to indicate senescent cells. Algorithms developed for this study analyzed the images, measured 87 different physical characteristics for each cell, and sorted the fibroblasts into groups.

Fibroblasts come in 11 different shapes and sizes, three of which are distinct to senescent skin cells, the researchers found. Only one subtype of senescent fibroblast, which the researchers named C10, was more prevalent in older donors.

In the petri dishes, each subtype responded differently when exposed to existing drug regimens designed to target and kill zombie cells. Dasatinib + Quercetin, a drug being tested in clinical trials, for example, most effectively killed C7 senescent fibroblasts but was limited in killing the age-associated C10 senescent fibroblasts.

Though further research is needed to verify which fibroblast subtype is harmful and which is helpful, the findings show that drugs can target one subtype and not the others.

“With our new findings, we have the tools ready to develop new drugs or therapies that preferentially target the senescence subtype that drives inflammation and disease as soon as it is identified,” Phillip said.

More precise targeting of senescence could benefit cancer treatments, the researchers said.

Certain therapies are being designed to trigger senescence in cancer cells, converting uncontrollably replicating cancer cells into dead-in-the-water zombie cells. While these therapies could stop tumor growth, they leave senescent cells in their wake. Conventional chemotherapies also push cells like fibroblasts toward senescence as a side effect. The buildup of senescent cells during treatment can be problematic as those cells may promote inflammation at a time when a patient’s immune system is at its most vulnerable.

Patients may benefit from a drug administered after chemotherapy that can sweep up the mess, removing harmful senescent cells while leaving behind the helpful senescent cells. These types of drugs are called senotherapies.

Next, the researchers plan to look at senescence subtypes in tissue samples, not just in flasks and petri dishes, to see how those subtypes might be associated with various skin diseases and age-associated diseases.

“We hope, with some more development, our technology will be used to help predict which drugs might work well for targeting senescent cells that contribute to specific diseases,” Phillip said. “Eventually, the dream is to be able to provide more information in a clinical setting to help with individual diagnoses and boost health outcomes.”

In their study, published in the American Journal of Preventive Medicine, the researchers found that when simple but strategic digital features, such as colour-coded nutritional quality signals and a healthier alternative prompt, were added to an online grocery shopping platform, the nutritional quality of shoppers’ carts improved significantly.

The team’s findings verified the effectiveness of deploying front-of-package (FOP) labels — which only marginally improve diet quality when used alone — alongside other interventions.

In a randomised trial conducted on NUSMart, an online grocery store designed by the Duke-NUS team, study participants were randomly divided into two groups and asked to make a total of three orders over a period of three to six weeks. While those assigned to the control group used a standard version of the store, those assigned to the experimental group used a version of the store with additional digital features, including:

Signalling nutritional quality with a traffic light: FOP labels resembling traffic light signals alerted shoppers to food products’ nutritional quality using three colour bands — green (best), amber and red with an “X” mark — for easy identification of foods to avoid.

Items were sorted into the colour bands based on their Nutri-Score (NS)[1] points, which were assigned according to energy, sugar, sodium, saturated fat, fruit/vegetable, protein, and dietary fibre levels per 100g/ml. The points were then converted into grades on a five-letter grading system, with A being the healthiest and E being the least healthy.

Sorting groceries by nutritional value: Using food products’ NS points, the researchers presented the items by order of nutritional value, with the healthiest options appearing first. Items in the control version of NUSMart appeared in alphabetical order.

Showing real-time cart feedback: Participants could track the nutritional quality of their grocery carts via a pie chart that indicated the proportion of items in each colour band. They could also compare their carts with a reference cart for health grocery shopping, which the researchers had curated using past data.

Suggesting healthier options: Shoppers could also view up to four healthier alternatives with similar prices and characteristics to each selected food product and replace their chosen product with the healthier alternative at the click of a button.

With these interventions, the nutritional quality of participants’ grocery carts improved from NS grade C to NS grade B, which was more significant than results from previous studies involving standalone FOP labels. The interventions also reduced the amount of calories (12.86 kcal), total fat (1.21g), saturated fat (0.85g), sugar (0.82g) and sodium (156.64mg) purchased.

Notably, the researchers found that shopping with the four digital features led to healthier food choices across all three orders.

Assistant Professor Soye Shin from Duke-NUS’ Health Services and Systems Research Programme, the study’s first author, said of the findings: “As online grocery shopping is rapidly gaining ground, we wanted to see if we could design low cost, scalable online tools that could be used to nudge consumers toward healthier choices at the point of purchase. These results show the potential of these tools to improve diet and health outcomes.”

These findings underscore the advantages of introducing diet quality labels, such as Singapore’s Nutri-Grade labelling initiative, which currently only includes beverages but is planned to expand to other food categories.

Senior author Professor Eric Finkelstein, from Duke-NUS’ Health Services and Systems Research Programme, said: “These results are encouraging but the next step is to work with retailers to incorporate these features into existing online stores. Only then will the full value of this approach be realised.”

Next, the researchers will expand the study to include consumers of low socio-economic status and little nutritional knowledge. They will also investigate if the multi-pronged intervention strategy has the potential to positively impact consumers’ health in the long term.

Professor Patrick Tan, Senior Vice-Dean for Research at Duke-NUS, commented on the study’s impact: “This study reflects how smart, evidence-based interventions — when applied at the right moment — can empower people to make better everyday choices for their health. It also shows how research can lead to practical tools that improve not just individual choices, but population health outcomes.”

Duke-NUS is at the forefront of biomedical research and translational innovations. This new study is part of the School’s ongoing efforts to improve global health through systems research and scientific breakthroughs.

While organisms use energy for a very wide variety of processes — from growth to cognition — one activity is a major drain for many animals: movement. For highly mobile animals, movement is as such a powerful lens through which to estimate energy usage.

Strong methods do exist for measuring animal movement in the context of energy expenditure, but these are limited by the physical size of the equipment used. Now, in a paper published in the Journal of Experimental Biology, researchers from the Marine Biophysics Unit at the Okinawa Institute of Science and Technology (OIST), in collaboration with Professor Amatzia Genin from the Hebrew University of Jerusalem, describe an innovative method for measuring energy usage during movement with video and 3D-tracking via deep learning. “The best method without space limitations has until now been ruled out in the study of about half of the world’s species, due to the reliance on wearable equipment,” says Dr. Kota Ishikawa, first author of the study. “With video, we now have a more inclusive method for studying energy usage in the context of animal behavior and ecology.”

Dynamic Body Acceleration (DBA) has long been the state-of-the-art method for estimating energy usage during movement. In essence, DBA involves measuring the oxygen consumption of a given species performing a given behavior in the lab, by simply measuring the amount of oxygen consumed through the activity. Oxygen is a good indicator of energy, as it is consumed as part of aerobic respiration to produce ATP — the ‘fuel’ providing energy for most bodily processes, including muscle contraction. The acceleration of the animal is measured simultaneously with an accelerometer, and in most cases, because the correlation between acceleration and oxygen consumption during the behavior is very strong, DBA provides a reliable estimate of energy consumption.

With the standard set in the lab, DBA is then measured in the wild, where reliably measuring oxygen consumption is impossible, through a wearable accelerometer. However, relying on physical equipment presents a major barrier. “To ensure accurate measurements without influencing the behavior during observation, researchers have used equipment that weighs at least ten times less than the animal. Given that the accelerometer and battery pack weigh 10-20 grams, this rules out the study of any animals below 100 grams — about half the world’s vertebrate species. It can also affect movement, especially when it depends on drag efficiency, such as swimming or flight,” says Dr. Ishikawa.

Their solution to this problem is elegantly simple. Instead of using a physical accelerometer to measure the movements, two cameras capture video footage of the behavior — in this case, a damselfish swimming in a fishtank — from multiple angles to reconstruct the behavior in 3D space. A few frames of the videos are then used to train a deep learning neural network to track the position of body features such as eyes, which allows researchers to subsequently measure the movement-related acceleration.

Both in the field and in the lab, once cameras are set up to capture animal movements, energy consumption can be estimated. For example, video-based DBA can be used in the context of collective behavior: “Energy expenditure during schooling of small fish has long remained mysterious,” explains Dr. Ishikawa. “For example, do the leading fish use more energy, and is schooling an energy-efficient form of movement? And what can that tell us about the ecology and evolution of fish schooling?”

With video-based DBA, the accurate measurement of energy usage during free-ranging animal behavior has been opened to the smaller half of the world’s vertebrate species, potentially enabling many new research avenues into the breadth of life on our planet.

A recent study, led by Professor Zi Yang MENG and co-authored by his PhD student Menghan SONG of HKU Department of Physics, in collaboration with researchers from the Chinese University of Hong Kong, Yale University, University of California, Santa Barbara, Ruhr-University Bochum and TU Dresden, has unravelled some of the secrets concealed within the entangled web of quantum systems.

Their findings, recently published in the journal Science Advances, push the boundaries of modern physics and offer a fresh perspective on how quantum matter operates at these enigmatic junctures. The study not only deepens our understanding of quantum mechanics but also paves the way for future discoveries that could revolutionise technology, materials science, and even our understanding of the cosmos.

What are Deconfined Quantum Critical Points?

In everyday life, we are familiar with phase transitions, such as water freezing into ice or boiling into steam. These transitions are well-understood and explained by thermodynamics. However, in the realm of quantum physics, phase transitions can occur at absolute zero temperature (-273.15 °C), driven not by thermal energy but by quantum fluctuations — tiny, unpredictable movements of particles at the smallest scales. These are known as quantum critical points.

Traditional quantum critical points act as boundaries between two distinct states: a symmetry-broken phase (ordered phase), where particles are neatly arranged, and a disordered phase, where particles are jumbled and chaotic. This kind of transition is well-described by the Landau theory, a framework that has been the foundation of our understanding of phase transitions for decades.

But deconfined quantum critical points (DQCPs) break this mould. Instead of a sharp boundary separating an ordered phase from a disordered phase, DQCPs lie between two different ordered phases, each with its own unique symmetry-breaking pattern, meaning the way particles are arranged or interact in one phase is fundamentally different from the other. This is unusual because, traditionally, phase transitions involve moving from an ordered state to a disorder one, not from one type of order to another. This distinction makes DQCPs fundamentally different and highly intriguing.

Scientists have debated for decades whether DQCPs represent continuous phase transitions (which are smooth and gradual) or first-order transitions (which are sudden and abrupt). Understanding DQCPs could provide new insights into how particles interact and how exotic states of matter emerge.

The Key to the Mystery: Entanglement Entropy

At the heart of this new study lies the concept of entanglement entropy, a measure of how particles in quantum systems are interrelated. It provides a way to quantify the amount of information shared between different parts of a system. Entanglement entropy offers a glimpse into the hidden structure of quantum systems, serving as a fundamental tool for probing quantum matter and understanding the nature of complex interactions that emerge at critical points.

Using advanced quantum Monte Carlo simulations (a computational method for modelling quantum systems) and rigorous theoretical analysis, researchers examine the behaviour of entanglement entropy in square-lattice SU(N) spin models — a theoretical framework designed to capture the essence of DQCPs.

Their meticulous computations revealed something extraordinary: at small value N (a parameter that determines the symmetry of the system), the behaviour of entanglement entropy deviated from expectations for smooth, continuous phase transitions. Instead, they found that DQCPs exhibit anomalous logarithmic behaviors, defyingthe theoretical constraints typically associated with continuous phase transitions.

The Breakthrough: A Critical Threshold and Conformal Fixed Points

One of the most striking revelations of the study was the identification of a critical threshold value of N. When N exceeds this threshold, DQCPs exhibit behaviours consistent with conformal fixed points — a mathematical framework that describes smooth, continuous phase transitions. This discovery is significant because it suggests that, under certain conditions, DQCPs can resemble continuous phase transitions. At these critical points, the system aligns with conformal fixed points, revealing a hidden structure in the quantum world where the boundaries between distinct phases dissolve, and matter exists in a state of extraordinary fluidity, defying the usual rules of physics.

Why This Matters

The implications of these findings are profound. DQCPs provide a unique testing ground for exploring the interplay of quantum mechanics, symmetry, and critical phenomena. Understanding their nature could unlock new insights into:

1. Exotic States of Matter: DQCPs are believed to be connected to the emergence of exotic phases, such as quantum spin liquids, which have potential applications in quantum computing and other advanced technologies.
2. Fundamental Physics: By challenging the traditional Landau paradigm, DQCPs force us to rethink the principles that govern phase transitions, potentially leading to new theoretical frameworks.
3. Technological Innovation: Insights gained from studying DQCPs could inform the design of novel materials with unique quantum properties, such as high-temperature superconductors or quantum magnets.

Conclusion The enigmatic world of deconfined quantum critical points stands at the frontier of modern physics, offering a glimpse into the uncharted territory of quantum mechanics. Through their meticulous investigation of entanglement entropy and SU(N) spin models, researchers have made significant strides in unravelling the mysteries of these critical phenomena.

This study was conducted in collaboration with Dr Jiarui ZHAO from the Chinese University of Hong Kong, Professor Meng CHENG from Yale University, Professor Cenke XU from the University of California, Santa Barbara, Professor Michael M. SCHERER from Ruhr-University Bochum, and Professor Lukas JANSSEN from TU Dresden.

Thyroid cancer is among the most prevalent cancers in Hong Kong and globally. Precision management of the disease often rely on two systems: (1) the 8th edition of the American Joint Committee on Cancer (AJCC) or Tumour-Node-Metastasis (TNM) cancer staging system to determine the cancer stage; and (2) the American Thyroid Association (ATA) risk classification system to categorise cancer risk. These systems are crucial for predicting patient survival and guiding treatment decisions. However, the manual integration of complex clinical information into these systems can be time-consuming and lack efficiency.

The research team developed an AI assistant that leverages large language models (LLMs), like ChatGPT and DeepSeek, which are designed to understand and process human language, to analyse clinical documents and enhance the accuracy and efficiency of thyroid cancer staging and risk classification.

The model leverages four offline open-source LLMs — Mistral (Mistral AI), Llama (Meta), Gemma (Google), and Qwen (Alibaba) — to analyse free-text clinical documents. The AI model was trained with a United States based open-access data with pathology reports of 50 thyroid cancer patients from The Cancer Genome Atlas Programme (TCGA), with subsequent validation against pathology reports from 289 TCGA patients and 35 pseudo cases created by endocrine surgeons.

By combining the output of all four LLMs, the team improved the overall performance of the AI model, achieving overall accuracy of 88.5% to 100% in ATA risk classification and 92.9% to 98.1% in AJCC cancer staging. Compared to traditional manual document reviews, this advancement is expected to halve the time clinicians spend on pre-consultation preparation.

Professor Joseph T Wu, Sir Kotewall Professor in Public Health and Managing Director of InnoHK D24H at HKUMed, emphasised the model’s remarkable performance. ‘Our model achieves more than 90% accuracy in classifying AJCC cancer stages and ATA risk category’, he said. ‘A significant advantage of this model is its offline capability, which would allow local deployment without the need to share or upload sensitive patient information, thereby providing maximum patient privacy.’

‘In view of the recent debut of DeepSeek, we conducted further comparative tests with a “zero-shot approach” against the latest versions of DeepSeek — R1 and V3 — as well as GPT-4o. We were pleased to find that our model performed on par with these powerful online LLMs,’ added Professor Wu.

Dr Matrix Fung Man-him, Clinical Assistant Professor and Chief of Endocrine Surgery, Department of Surgery, School of Clinical Medicine, HKUMed, stated, ‘In addition to providing high accuracy in extracting and analysing information from complex pathology reports, operation records and clinical notes, our AI model also dramatically reduces doctors’ preparation time by almost half compared to human interpretation. It could simultaneously provide cancer staging and clinical risk stratification based on two internationally recognised clinical systems.’

‘The AI model is versatile and could be readily integrated into various settings in the public and private sectors, and both local and international healthcare and research institutes,’ said Dr Fung. ‘We are optimistic that the real-world implementation of this AI model could enhance the efficiency of frontline clinicians and improve the quality of care. In addition, doctors will have more time to counsel with their patients.’

‘In line with government’s strong advocacy of AI adoption in healthcare, as exemplified by the recent launch of LLM-based medical report writing system in the Hospital Authority, our next step is to evaluate the performance of this AI assistant with a large amount of real-world patient data. Once validated, the AI model can be readily deployed in real clinical settings and hospitals to help clinicians improve operational and treatment efficiency,’ explained Dr Carlos Wong, Honorary Associate Professor in the Department of Family Medicine and Primary Care, School of Clinical Medicine, HKUMed.

The study was led by Professor Joseph Wu Tsz-kei, Sir Robert Kotewall Professor in Public Health in the School of Public Health, and Managing Director & Lead Scientist of InnoHK D24H; Dr Matrix Fung Man-him, Clinical Assistant Professor and Chief of Endocrine Surgery in the Department of Surgery, School of Clinical Medicine; and Dr Carlos Wong King-ho, Honorary Associate Professor in the Department of Family Medicine and Primary Care, School of Clinical Medicine, and Senior Research Director in InnoHK D24H; all under HKUMed. The first authors were Dr Eric Tang Ho-man and Dr Tingting Wu from InnoHK D24H.