Making wheat more resilient to climate change without compromising yields has become an urgent priority for the agricultural sector. Now, a study led by a research team from the University of Barcelona and the Agrotecnio research center has identified an innovative way to address this challenge: combining advanced technology and artificial intelligence to select the best varieties of this crop.

The study, published in the journal Plant Phenomics, suggests a shift in perspective: it is necessary to focus not only on yield, but also on wheat’s ability to maintain consistent harvests despite changing weather conditions. The findings indicate that this combination of productivity and stability is key to ensuring safe harvests under variable environmental conditions.

The authors of the study are researchers Jara Jauregui, José Luis Araus and Shawn Carlisle Kefauver, from the Department of Evolutionary Biology, Ecology and Environmental Sciences at the UB’s Faculty of Biology and Agrotecnio; Nieves Aparicio and Sara Álvarez, from the Agro-technological Institute of Castilla y León (ITACyL), and María Teresa Nieto, from the National Institute for Agricultural and Food Research and Technology (INIA-CSIC).

Drones for monitoring wheat crops

The team analyzed 64 varieties of durum wheat grown under two different Mediterranean conditions: irrigated and rain-fed. The aim was to identify which genotypes combine high yields with a stable performance across variable environments, with differences in temperature and water availability.

One of the most surprising findings is that the selected varieties are not those that retain their green leaves the longest until the end of the season, but rather those that grow vigorously at the start and mature slightly earlier.

In contrast, the rejected lines showed low initial vigor and retained their green leaves for longer, which does not guarantee a better yield.

As part of the project, the team used ground sensors and drones equipped with RGB, multispectral and thermal cameras, enabling them to monitor crop development throughout the entire growing cycle. This technology provides key information about the wheat before harvesting, eliminating the need for harvesting and reducing both the costs and the time required for analysis.

Using all this data, the team trained artificial intelligence models capable of predicting both the yield and the stability of production for the different varieties with a high degree of accuracy.

This strategy could be a very useful tool for plant breeding programs and could help develop wheat varieties that are equipped to meet the challenges of climate change.

Greener doesn’t always mean better

The researchers first analyzed, separately, the yield and stability traits of durum wheat. They found that the genotypes with the highest yields are characterized by high initial vigor and sustained greenness during the rapid growth phases up to the end of the growing season. In contrast, the most stable genotypes exhibit lower initial vigor, slower growth and a shorter cycle, enabling them to make better use of the resources available for grain production. To identify a balance between these compensatory mechanisms, the experts developed a varied selection method that combines competitive yield with good stability.

The study concludes that vigorous early growth combined with early maturation is a key factor to achieving more consistent yields under variable environmental conditions, helping wheat cope better with drought and high temperatures.

Professor Gyu Rie Lee of the Department of Biological Sciences successfully designed artificial proteins that selectively recognize specific compounds using AI through joint research with Professor David Baker. The research, published in the journal Nature Communications, is characterized by using AI to design proteins that recognize specific compounds from scratch (de novo) and implementing them as functional biosensors.

While the conventional approach mainly involved searching for natural proteins or modifying some of their functions, this research is highly significant in that it “custom-built” proteins with desired functions through AI-based design and even completed experimental verification.

In particular, the research team successfully designed a protein that selectively recognizes the stress hormone cortisol and implemented an AI-designed biosensor based on it. This is evaluated as a case that extends beyond protein design to actual measurable sensor technology, solving the long-standing challenge of small-molecule recognition in the field of protein design.

These research results have applications in fields including disease diagnosis, new drug development and environmental monitoring. The technology can precisely detect biomarkers in the blood to diagnose diseases early and contribute to the development of targeted therapies through the design of proteins that selectively recognize specific molecules.

Furthermore, it is expected that the implementation of customized biosensor technology will become possible, such as real-time monitoring of air and water quality through the development of sensors that detect environmental pollutants.

Designing new proteins (de novo proteins) that recognize compounds has been considered a challenge in the field of protein design for a long time because it requires precise calculations at the atomic level. The research team developed an AI model that precisely reflects protein-ligand interactions and successfully designed binding proteins using it.

As a result, artificial binding proteins were designed for six types of compounds, including metabolites and small-molecule drugs, and their functions were verified through experiments. In particular, a cortisol biosensor was developed by designing a chemical-induced dimer based on a new protein that binds with cortisol.

A provisional patent for the relevant design technology has been filed in the United States.

Professor Gyu Rie Lee stated, “This research experimentally proves that AI can be used to design proteins that precisely recognize specific compounds,” and added, “We plan to expand this into protein design technology that can be utilized in various fields such as disease diagnosis, new drug development, and environmental monitoring in the future.”

Professor Gyu Rie Lee of the KAIST Department of Biological Sciences participated in this research as the first author, and Professor David Baker as the corresponding author.

Director Do-Heon Lee, a mentor professor of the AI-CRED Innovative Drug Research Group, said, “This achievement is a meaningful result derived through cooperation between InnoCORE researchers and a global scholar.

“We will further strengthen our research capabilities based on active research collaboration with postdoctoral researchers recruited through the InnoCORE project to continue creating innovative results in the AI drug development and bio-fields.”

The KAIST InnoCORE Research Group aims to accelerate AI-based scientific and technological innovation and promote global joint research by supporting top-tier domestic and international postdoctoral researchers to devote themselves to the development of AI convergence technology in a cutting-edge collective research environment.

Researchers at the Donald Danforth Plant Science Center have solved a long-standing mystery of how a model green microalga reorganizes its central metabolism to supercharge growth when given access to both light and a carbon source—a finding with broad implications for developing algae as a sustainable source of renewable fuels, bioproducts, and biomass. Their study is published in the Proceedings of the National Academy of Sciences.

Algae are among Earth’s most productive life forms. Along with other photosynthetic microbes, they capture roughly half of all carbon absorbed from the atmosphere globally each year. Unlike land crops, algae can grow on land unsuitable for food production and generate oil, protein, and other valuable compounds at rates 10 to 50 times greater than most terrestrial plants. Yet translating that productivity into reliable, scalable bioproduct yields has proven difficult—largely because the metabolism governing how algae partition carbon under different growth conditions was poorly understood.

Using isotope-assisted metabolic flux analysis (MFA)—a technique that measures how carbon moves through cellular pathways—the research team led by Somnath Koley, previously in the Allen lab at the Danforth Center, compared algae grown on light alone versus algae grown on both light and acetate, a condition known as mixotrophic growth.

Rather than simply combining the two energy sources, the cells were found to fundamentally rewire their metabolism: activating highly efficient biochemical pathways to conserve carbon, suppressing a separate carbon-costly process, and strategically reducing photosynthesis in ways that lowered the burden of protein production and enabled faster growth (which greatly surpassed the additive effect of light-only growth plus acetate-only growth)—a result that prior computational models had failed to predict.

“What metabolic flux analysis reveals is the actual operating strategy of the cell—not a snapshot of gene or protein levels from -omics data, but the real rates at which carbon is moving through each pathway,” said Doug K. Allen, Ph.D., principal investigator at the Danforth Center and one of the corresponding authors. “Those real-world flux constraints are what you have to work with if you want to rationally engineer algae for higher yields.”

“Without flux analysis, we couldn’t have resolved the long-standing paradox of how acetate affects algae growth,” said James G. Umen, Ph.D., principal investigator at the Danforth Center and another corresponding author. “This study shows that metabolism is fundamentally different—and far more efficient—when light and acetate are both present, and that insight is critical for anyone trying to engineer algae for higher productivity.”

“Doug and Jim’s teams have provided something rare and valuable: a quantitative map of how algae actually manage carbon. That’s the kind of foundational insight that accelerates the path from discovery to real-world solutions—whether that’s a more sustainable fuel, a new biomass crop, or a bioproduct that reduces our dependence on fossil fuels,” said Giles Oldroyd, Ph.D., President of the Danforth Center.

A study published in Cell Research advances a central idea in stem cell biology by identifying a checkpoint that controls the identity of many different types of stem cells across developmental stages. For nearly two decades, scientists have understood that stem cell self-renewal depends on blocking differentiation signals—a concept described in earlier work, including Qi-Long Ying and Austin Smith’s 2008 Nature paper titled “The ground state of embryonic stem cell self-renewal.”

Now, researchers from the labs of Ying at USC and Guang Hu at the National Institute of Environmental Health Sciences (NIEHS), one of the National Institutes of Health (NIH), have identified the protein GSK3α as a “stemness checkpoint” that drives differentiation and that can be inhibited to maintain stem cell identity.

This discovery introduces a new conceptual framework: Rather than viewing stem cell maintenance as the result of many unrelated signaling conditions, distinct stem cell types share common checkpoints.

“We already knew that blocking differentiation is essential for maintaining stem cells,” said co-corresponding author Ying, professor of stem cell biology and regenerative medicine at the Keck School of Medicine of USC. “What this study shows is that there are specific checkpoints controlling this process, and that these checkpoints are shared across different stem cell states.”

This framework could inform the development of better conditions for maintaining stem cells in the laboratory, which is crucial for providing renewable sources of cells for studying development, modeling disease, testing drugs, developing cell therapies and regenerating tissues.

A common checkpoint

For most of the experiments, first authors Duo Wang from USC and Xiukun Wang from the NIEHS and their colleagues studied mouse stem cells derived at two distinct developmental stages: embryonic stem cells (mESCs) and epiblast stem cells (mEpiSCs). These cell types normally require very different laboratory conditions to maintain their identities.

Despite these differences, both cell types responded to GSK3α as a common checkpoint. By inhibiting GSK3α, the team demonstrated that mESCs and mEpiSCs multiplied to maintain stable self-renewal and preserved their identities, even when grown together in the same dish for more than a month.

The researchers extended these findings to additional stem cell types, including neural stem cells and “formative stem cells,” which represent an intermediate state between mESCs and mEpiSCs, demonstrating that this checkpoint mechanism operates broadly across stem cell states.

Importantly, complementary studies showed that GSK3α serves as a stemness checkpoint across species, including rats, rabbits, cows and humans, highlighting its fundamental biological role.

In addition to helping scientists develop better conditions for maintaining and expanding stem cells in the laboratory, the findings may also have implications for aging.

“This study suggests that stem cell aging may, in part, reflect the progressive activation of differentiation checkpoints,” said Ying. “Controlling these checkpoints could provide a new strategy for maintaining tissue health over time.”

Co-corresponding author Hu added, “More broadly, the work establishes a new framework for understanding stem cell regulation across development and disease, with potential applications in regenerative medicine, disease modeling and cancer research.”

Stem cells are the body’s ultimate shape-shifters, sustaining tissues by balancing two competing demands: maintaining their own population and generating specialized descendants. In many tissues, some early descendants can revert to a stem cell state through a process known as dedifferentiation. This ability can help replenish the stem cell pool when stem cells are lost.

Probing what makes stem cells unique

In a new study published in PNAS, researchers at Whitehead Institute identify two complementary mechanisms that allow cells to preserve stem cell potential while adopting distinct identities.

Led by Whitehead Institute Member Yukiko Yamashita and Yamashita Lab postdoc Amelie Raz, the study focuses on the male fruit fly germline stem cells, which give rise to sperm. These cells sit at the foundation of a lineage that continues across generations.

To understand what distinguishes these stem cells, the researchers analyzed RNA, the intermediary molecules that link genes in DNA to the proteins they encode. RNA quantities typically reflect which genes a cell is using—which in turn reflects a cell’s identity.

The researchers expected to find a set of RNAs unique to stem cells. Instead, they discovered that stem cells and their immediate descendants share seemingly identical RNA profiles.

“We didn’t have anything that was specific to stem cells,” Raz says. “It turned out that that was actually the key to understanding how you make them.”

The difference between these cell types lies not only in which RNAs are present, but in whether the cells are still making them. Stem cells continue producing these RNAs, while their descendants inherit many of the same molecules but stop making new copies of RNA.

This means RNA alone does not fully define a cell’s state. In these descendant cells, the shared RNAs reflect an earlier state, not the same productive gene program seen in stem cells.

“On the level of RNA, they’re the same,” Raz says. “But they’re different in what’s actually happening in the nucleus—whether that RNA is being actively produced.”

Independent signals shape cell fates

The study also clarifies how signals from the surrounding environment help determine what path a cell follows. Stem cells reside in a specialized microenvironment known as a niche, which sends molecular cues that influence cell behavior. Two well-studied signaling pathways—Bmp and Jak-Stat—have long been known to regulate germline stem cells.

Previous models assumed these pathways worked together or redundantly. However, the new findings show that they instead act independently, each controlling a different subset of genes.

“What we found was that they’re acting on completely separate parts of this gene activity program,” Raz says.

Because the pathways operate independently, their combined activity defines distinct cellular states. When both signals are active, cells maintain stem cell identity. When neither is active, cells continue along a differentiation pathway. When only one pathway is active, cells can revert toward a stem cell state through dedifferentiation.

This modular arrangement allows cells with the same underlying potential to follow different paths depending on the signals they receive.

The findings help explain why many stem cell populations rely on multiple signaling pathways. Rather than serving as backups for one another, these pathways can regulate different parts of cell behavior and work together to shape a cell’s trajectory.

“In many stem cell populations, multiple signals have been thought to be redundant,” says Yamashita, who is also a professor of biology at MIT and an HHMI Investigator. “Here, we show that they can have distinct roles to determine whether a cell self-renews, differentiates, or reverts in combination.”

Rethinking stem cell regulation models

More broadly, the work shows that knowing which molecules are present in a cell does not always reveal how that cell is functioning. Two cells can appear identical by standard molecular measures even when they are operating in different regulatory states.

The study also lays the groundwork for future research. Raz and colleagues have identified a set of genes linked to this early germline state in fruit flies and are now investigating what those genes do and how they help govern stem cell behavior.

“Now that we know what’s there, the next step is understanding what those RNA molecules are doing,” Raz says.

Additionally, the work suggests that long-standing models of stem cell regulation may be incomplete, even in systems that have been studied for decades.

“What we are showing is that these pathways aren’t necessarily working in the way people had assumed,” Raz says. “There’s almost certainly more to it.”

Researchers at the National University of Singapore (NUS) have built a molecular “leash” to pull directly on a force-sensing protein called Piezo1, and discovered it switches on at about 15 piconewtons, proving that it can be activated by physical tethers, not only by membrane deformation. The study is published in the journal Nature Sensors.

The team developed a new method to directly measure how tiny mechanical forces activate a key protein in our cells, known as Piezo1, a sensor that allows cells to “feel” physical forces. Using a DNA-based approach, the team was able to apply extremely small, well-controlled forces on the order of trillionths of a newton to individual Piezo1 channels. At the same time, they monitored the channel’s activity in real time using a fluorescent signal that lights up when calcium ions enter the cell. Their key finding: Piezo1 can be switched on by a force of about 15 piconewtons, providing the first precise measurement of the force needed to activate this important sensor.

Why this matters

Cells constantly experience physical forces, from blood flow and touch to tissue movement. Proteins like Piezo1 help cells detect and respond to these forces, playing roles in processes such as blood pressure regulation, immune responses, and tissue repair and development. However, until now, scientists could not directly measure the amount of force required to activate Piezo1, because existing techniques also changed the shape of the cell membrane, making results difficult to interpret.

A new way to apply force using DNA

The NUS team, led by Professor Liu Xiaogang from the Department of Chemistry, NUS and Professor Yan Jie from the Department of Physics, NUS solved this problem by attaching Piezo1 to tiny beads using strands of DNA. This setup enabled the researchers to apply precise, well-calibrated forces directly to the protein while avoiding unwanted effects from membrane stretching. It also allowed them to measure the protein’s response at the single-molecule level.

A shift in understanding

Previously, scientists believed Piezo1 mainly responds to forces acting through the cell membrane. This study shows that forces transmitted through structures like the extracellular matrix or cytoskeleton can also directly activate the channel. This provides strong evidence for an alternative mechanism, sometimes described as “force-from-filament,” and suggests that cells may sense mechanical signals in more ways than previously thought.

Using this highly precise and flexible DNA-based system, the researchers showed that Piezo1 can be activated in a controlled, repeatable, and reversible way by applying defined forces. The platform offers exceptional spatial control, allowing scientists to study the response of individual ion channels to mechanical forces with high accuracy. Importantly, this method is not limited to Piezo1, as it can be adapted to investigate other force-sensitive proteins, offering a versatile tool to better understand how mechanical forces influence biological processes.

Dr. Sui Mingyu, who is part of the research team, said, “This work represents both a fundamental and technological advance in mechanobiology. By establishing a clear and quantitative link between applied force and ion channel activation while separating this effect from changes in the cell membrane, this approach opens up new ways to study the response of cells to physical forces in both health and disease.”

In the future, this platform could help uncover force-dependent signaling pathways, support the development of new mechanosensitive therapies, and enable the design of materials that respond to mechanical stimuli.

A new tool makes it possible to screen millions of tiny protein fragments and select those that can be recognized by the immune system. The CIC biomaGUNE Center for Cooperative Research in Biomaterials has developed epiGPTope, a system that uses machine learning to generate and classify epitopes, in collaboration with the company Multiverse Computing.

The immune system is triggered by the presence of viruses or bacteria. When the antibodies produced recognize the epitopes, a small part of these viruses or bacteria, they launch an attack strategy. These epitopes are small fragments of protein recognized by antibodies or by immune cell receptors. So discovering new epitope sequences that target specific antibodies is essential for the development of diagnostic tools, immunotherapies and vaccines.

CIC biomaGUNE’s Biomolecular Nanotechnology laboratory, led by the Ikerbasque Research Professor Aitziber L. Cortajarena, is creating a library or database of hundreds of thousands of synthetic epitopes using this AI-based technique. The work is published in the journal ACS Synthetic Biology.

This method for creating biologically viable sequences enables the research group to generate and select synthetic epitopes more quickly and cost-effectively, as well as to classify them according to their viral or bacterial origin, thereby facilitating their application in biotechnology and biomedicine.

“From among millions of possible combinations, we identify synthetic epitopes that are very similar to natural epitopes, and which can be recognized by antibodies,” explained Aitor Manteca, a research associate in the group. The aim is “to see what applications these molecules might have in medical research, drug development and biotechnology.”

“What is more, we are able to determine whether an epitope comes from a bacterium or a virus.” That way, “a rational library of epitopes comprising hundreds of thousands of units rather than hundreds of millions, which are stored in the laboratory” can be built. “These are physical collections of molecules for experimentation purposes,” he added.

Devices for point-of-care diagnosis

However, these epitopes do not remain confined to small test tubes. Real-world applications are being sought for them. Once this initial screening has been carried out, the protein fragments are analyzed using microfluidic systems.

This is technology “that allows a single epitope to be tested against a specific antibody in a highly precise, rapid and cost-effective manner, yielding numerous results in a short space of time,” explained Manteca.

Microfluidics allows the experiments to be carried out on tiny droplets, which act as individual reactors, thus using very small quantities of molecules. “It is possible to analyze millions of different combinations simultaneously within a short period of time,” added the CIC biomaGUNE researcher.

“So it is possible to find out in advance “which sequences will trigger an immune response and drive forward, for example, the development of diagnostic techniques and new point-of-care devices capable of measuring the presence of a bacterium or virus in the body, blood, water, and so on,” explained Dr. Aitor Manteca.

These developments are particularly relevant for transfer to the industrial sector, as in the case of the enterprise Taldeki Biosolutions, which uses a detection technology licensed by CIC biomaGUNE. In this context, the capacity to rapidly and rationally generate and select epitopes means that the identification and validation of recognition elements to be integrated into new sensors can be significantly sped up.

This approach is expected to have a direct impact on the development of advanced diagnostic solutions, because the rapid development and selection of epitopes will enable the scope of these sensor technologies to be significantly expanded to cover a wide range of detection applications.

This encompasses not only the biomedical field, but also environmental and biotechnological contexts by strengthening the link between fundamental research and industrial application.

The use of machine learning algorithms and artificial intelligence is revolutionizing every field in which it is applied. Biotechnology is no exception; in fact, one could say that bioinformatics is one of the first fields to adopt many of the new technologies currently being developed.

CIC biomaGUNE is streamlining the building of a collection of hundreds of thousands of molecules and the study of their potential applications in medicine, pharmacology and biotechnology.

Researchers at Cornell University have developed a powerful new genetic toolkit that allows scientists to study how genes function at the level of individual cells, an advance that could accelerate discoveries in development, neuroscience, and disease. The work is published in the journal eLife.

The system builds on MAGIC (Mosaic Analysis by gRNA-Induced Crossing-over), a method originally created by the labs of Chun Han, associate professor in the Department of Molecular Biology and Genetics in the College of Agriculture and Life Sciences (CALS) and the Weill Institute for Cell and Molecular Biology. MAGIC uses CRISPR gene editing to generate individual mutant cells within otherwise normal tissue, enabling precise comparisons within a living organism.

In the new study, graduate researcher Yifan Shen expanded the approach into a genome-wide toolkit for Drosophila melanogaster, creating resources that work across all chromosomes and allow researchers to study genes that were previously difficult, or impossible, to analyze at single-cell resolution.

“It saves at least a couple of months for studying a single mutation compared to traditional methods,” Han said. “If someone wants to screen hundreds or thousands of mutations, the time saved will be years or more.”

Traditional mosaic analysis often requires researchers to spend weeks or months recombining genetic elements before experiments can begin. By contrast, the expanded MAGIC system works directly with existing genetic stocks, significantly lowering technical barriers while relying on standard laboratory equipment.

The toolkit also introduces improved fluorescent markers that make mutant cells brighter and easier to track under a microscope. In earlier versions, visualizing fine cellular structures could be challenging.

“Our final design beautifully illuminated the whole neurons, down to the finest branches,” Han said. “This was a huge help for analyzing dendrite morphology of individual neurons.”

With genome-wide coverage, the system opens the door to large-scale genetic screens at single-cell resolution. Researchers can now pair MAGIC with existing resources such as Drosophila deficiency libraries to systematically scan the genome for genes involved in key biological processes.

“Many fundamental biological processes are still poorly understood due to the previous inability to screen all genes at the individual cell level,” Han said. “The combination of deficiency libraries and the MAGIC kit greatly accelerate the gene discovery process.”

The researchers also demonstrated that the system works in hybrid animals derived from different Drosophila species, an application that had been difficult or impossible with previous tools.

“This technique should allow researchers to ask some very interesting questions about speciation in ways impossible before,” Han said.

In addition, the team overcame longstanding technical challenges associated with the fruit fly’s fourth chromosome, a region that has historically been difficult to study. The new toolkit enables more reliable analysis of genes on that chromosome, potentially revealing previously overlooked biological functions.

To maximize accessibility, the researchers have made the toolkit broadly available through community repositories, allowing other labs to adopt the system without specialized training.

“We have benefited so much from the supportive Drosophila community,” Han said. “If other labs can now use our system to study their most important questions in ways not possible before, we feel that we have reached our goals.”

All fly stocks generated in this research have been deposited to the Bloomington Drosophila Stock Center, and the molecular tools are available through Addgene, ensuring broad access for the research community.

Cell cultures—single layers of cells grown in a small dish—have enabled researchers to study biological growth, develop or test drugs and even discover what causes some diseases. Cell spheroids, 3D versions of cell cultures built using a process known as cell aggregation, are the next step in advancing this work, capable of more closely modeling real tissue. A new technology, invented by researchers from Penn State and detailed in a paper published in Advanced Science, could breathe fresh air into bottom-up tissue fabrication and potentially large-scale tissue engineering by addressing these issues.

According to Amir Sheikhi, an associate professor at Penn State, there are several major flaws that limit the survivability and function of the cells comprising spheroids.

First, oxygen and nutrients cannot breach the outer layer to keep the inner cells functioning; second, without the help of blood vessels or channels to remove cellular waste or deliver nutrients and oxygen, it is difficult to grow spheroids to be large enough for use in tissue engineering; and third, these spheroids do not initially have an extracellular matrix, which provides critical structural and chemical support to the cells.

The team’s biohybrid spheroids—a mixture of living cells and microgels, tiny materials that mimic the supportive tissue that surrounds cells in the body—can rapidly self-assemble or “snowball” in size, while still allowing cell-supporting oxygen and nutrients to reach the cells inside.

Sheikhi, the Dorothy Foehr Huck and J. Lloyd Huck Early Career Chair in Biomaterials and Regenerative Engineering, said the team plans to further develop biohybrid spheroids capable of mimicking the properties of tissue, with the eventual goal of enabling commercial-scale tissue engineering and organ biofabrication.