Bacterial cellulose promotes plant tissue regeneration, study shows

A study has elucidated the mechanism by which bacterial cellulose mediates plant tissue regeneration. The work has been published today in the journal Science Advances and includes collaborations with researchers of the Institute of Materials Science of Barcelona (ICMAB-CSIC) and Colorado State University.

Bacterial cellulose (BC), synthesized by certain bacteria as a biofilm, consists of highly pure cellulose fibers. BC has been widely used in human biomedical applications showing a high degree of biocompatibility, but its potential healing effects in plants were unknown.

In this work, scientists have demonstrated that BC patches induce plant tissue regeneration and have identified for the first time the precise molecular mechanism underlying the process. Wounded leaves of the model plants Nicotiana benthamiana and Arabidopsis thaliana were covered with BC patches and formation of new cells on both sides of the cut was observed two days post-wounding, reaching complete wound closure after seven days.

This wound healing process was promoted by BC but not by other structurally similar matrixes such as plant cellulose, indicating that BC had specific features beyond preventing dehydration or promoting physical contact.

Scientists discovered that BC patches contain cytokinins, a class of hormones involved in plant development.

“Plants with defective cytokinin signaling did not respond to BC patches, confirming that cytokinins are crucial for triggering regeneration,” explains Nerea Ruiz-Solaní, a co-first author of the study.

The team also identified production of oxidative stress (ROS) at the wound sites upon BC application. Genomic and bioinformatic analyses enabled scientists to identify the specific genes involved in the process, which are typically associated with biotic responses, i.e. defense mechanisms against pathogens. The transcription factor WRKY8, which regulates defense responses, was found to interact directly with the promoter of GSTF7 gene leading to ROS accumulation.

Importantly, it is the concurrent activation of both the cytokinin and defense signaling pathways that results in the observed tissue regeneration, a novel finding since these mechanisms were previously studied independently. Further research would be needed to fully elucidate the pathways leading to cell cycle activation and functional differentiation during regeneration.

“Bacterial cellulose triggers a unique transcriptional program that differs from typical wounding-and callus-induced regeneration,” states Miguel Moreno-Risueño, co-leader of the study and an expert in plant regeneration at the CBGP.

This research holds significant implications for agricultural practices, including wound healing to prevent infections, and applications in grafting, pruning, and ornamental plant care, particularly in vineyards, rose cultivation, and stone pine production.

This work started a few years ago, back in 2016, with joint projects between CRAG and ICMAB-CSIC research groups led by Núria Sánchez Coll and Anna Laromaine, respectively, and in collaboration with companies such as AGROMILLORA and Forestal Catalana S.A. (Plant Healing and Plant Nanohealing projects), which conducted preliminary field trials.

However, further field studies are still needed to confirm the efficacy of BC patches in grafting. More technology transfer resources are needed to bridge the gap between fundamental research and the productive sector, with significant economic implications.

Núria Sanchez Coll, CSIC researcher at CRAG and co-leading author, highlights the collaborative nature of this research, saying, “This work has fostered very interesting collaborations with other research groups as well as industry, paving the way for further studies on plant regeneration mechanisms and potential biotechnological applications.”

Acidification kills H5N1 in waste milk, reducing risk of bird flu

Pasteurization is the only widely recognized method of killing H5N1, the virus that causes bird flu, in milk. However, pasteurization can be expensive and fewer than 50% of large dairy farms pasteurize waste milk.

Waste milk includes colostrum, the first milk after calving; milk from cows treated with antibiotics or other drugs; or any other factor that can make milk unsuitable and unsellable for human consumption. On farms, raw waste milk poses a potential risk of spreading avian flu, which so far has been confirmed in dairy cattle in 16 states.

University of California, Davis, researchers have found that acidification can kill H5N1 in waste milk, providing dairy farmers an affordable, easy-to-use alternative to pasteurization of waste milk. The Journal of Dairy Science published the study.

“There can be a quite significant cost to have pasteurization as an option on the farm,” said co-corresponding author and veterinary epidemiologist Richard Van Vleck Pereira, with the UC Davis School of Veterinary Medicine. “In our laboratory tests, we found that acidifying milk to a pH of 4.1 to 4.2 with citric acid effectively deactivates the virus.”

The UC Davis research team will next conduct on-farm testing of milk acidification in waste milk containing H5N1. They will develop practical guidelines for farmers to implement acidification of waste milk as a protocol on the farm.

A sustainable solution

Pereira said citric acid is inexpensive. Acidified waste milk is also safe to be used to feed pre-weaned calves. The acidification process takes only six hours to fully kill the virus and doesn’t require refrigeration, further reducing costs and increasing safety of farm workers handling milk.

Hobby farmers milking one or two cows or large commercial dairy farms could implement milk acidification without having to invest in large equipment.

“When we started this project, we were carefully thinking about not just deactivating the virus but developing a method that could be affordable, accessible and sustainable for farmers to use,” he said.

Some U.S. dairy farms already practice milk acidification. Lowering milk pH to a level unsuitable for bacterial growth can kill bad bugs and prevent contamination without causing health issues in calves.

“We believe acidification is a novel and effective way to contain the spread of H5N1 on dairy farms and help protect livestock, pets and people,” Periera said.

Zika Virus Creates Tunnels to Cross Placenta in Stealth Mode and Infect Fetus

Infection with Zika virus (ZIKV) in pregnancy can lead to neurological disorders, fetal abnormalities, and fetal death. Until now it’s not been clear how the virus manages to cross the placenta, which forms a strong barrier against microbes and chemicals that could harm the fetus. The results of a laboratory study carried out by researchers at Baylor College of Medicine, in collaboration with a team at Pennsylvania State University, have identified a strategy that Zika virus uses to covertly spread in placental cells, raising little alarm in the immune system. The team suggests their findings could point to new therapeutic strategies against the virus.

“The Zika virus, which is transmitted by mosquitoes, triggered an epidemic in the Americas that began in 2015 and by 2018 had reached as many as 30 million cases,” said Indira Mysorekar, PhD, E.I. Wagner Endowed, M.D., Chair Internal Medicine II, chief of basic and translational research and professor of medicine–infectious diseases at Baylor. “Understanding how Zika virus spreads through the human placenta and reaches the fetus is critical to prevent or control this devastating condition.

Zika virus is a mosquito-borne virus in the Flaviviridae family, and infection can lead to neurological disorders and fetal abnormalities such as microcephaly, and fetal death, “collectively known as congenital Zika syndrome,” the authors explained. “The propensity for horizontal and vertical transmission, and the ability to traverse blood-tissue barriers, including the blood-placental barrier, of ZIKV are unique among Flaviviridae.” The researchers noted that their own studies in mice, and work by others, have shown that ZIKV can infect fetal trophoblasts and endothelial cells of the placenta, which form the primary barrier between the maternal and fetal circulations. By infecting these cells the virus can enter the fetal circulation.

The researchers’ newly reported laboratory study has now discovered that Zika virus builds underground tunnels, a series of tiny tubes called tunneling nanotubes TNTs, in the placental trophoblasts, which facilitate the transfer of viral particles to neighboring uninfected cells. This ability is reliant on a viral protein NS1. “TNT formation is driven exclusively via ZIKV non-structural protein 1 (NS1),” they wrote.

“Zika is the only virus in its family, which includes dengue and West Nile viruses among others, whose NS1 protein triggers the formation of tunnels in multiple cell types,” Michita said. “Other viruses unrelated to Zika, such as HIV, herpes, influenza A, and SARS-CoV-2, the virus that causes COVID-19, also can induce tiny tunnels in cells they infect and use the tunnels to spread to uninfected cells. This is the first time that tunneling has been shown by Zika virus infection in placental cells.”

Interestingly, the tiny conduits provided a means to transport not only viral particles, but also RNA, proteins, and mitochondria, a cell’s main source of energy, from infected to neighboring cells. “We demonstrate that ZIKV infection or NS1 expression induces elevated mitochondria levels in trophoblasts and that mitochondria are siphoned via TNTs from healthy to ZIKV-infected cells,” the team wrote. Added co-author Long B. Tran, a graduate student in the Mysorekar lab, “We propose that transporting mitochondria through the tunnels may provide an energetic boost to virus-infected cells to promote viral replication.”

The study findings showed how TNT-mediated trafficking also allows Zika cell-to-cell transmission that is “camouflaged from host defenses.” Michita further commented, “We also show that traveling through the tiny tunnels can potentially help Zika virus avoid the activation of large-scale antiviral responses, such as interferon lambda (IFN-lambda) defenses implemented by the placenta. Mutant Zika viruses that do not make tiny tunnels induce robust antiviral IFN-lambda response that can potentially limit the spread of the virus.”

Mysorekar continued, “Altogether, we show that Zika virus uses a tunneling strategy to covertly spread the infection in the placenta while hijacking mitochondria to augment its propagation and survival. We propose that this strategy also protects the virus from the immune response. These findings offer vital insights that could be used to develop therapeutic strategies targeted against this stealth transmission mode.”

In their paper, the team concluded, “Our investigation reveals a previously unknown mechanism of intercellular transmission exploited by ZIKV, setting it apart from other orthoflaviviruses  … Together our findings identify a stealth mechanism that ZIKV employs for intercellular spread among placental trophoblasts, evasion of antiviral interferon response, and the hijacking of mitochondria to augment its propagation and survival and offers a basis for novel therapeutic developments targeting these interactions to limit ZIKV dissemination.”

The team acknowledged that further research will be needed to investigate the molecular mechanism by which ZIKV NS1 induces TNTs, and whether monoclonal antibodies or NS1-based vaccines target TNT formation in ZIKV-infected cells, to potentially limit viral infection and spread.

From farmland to construction: Bacteria strains offer sustainable biocement solution

A recent study examines the effectiveness of environmental strains for the production of biocement. The study’s lead author, Dimitrios Terzis, is an EPFL senior scientist and a co-founder of Medusoil, a company that produces organic binders and that opened a production plant in Vaud in 2024.

“For me, it’s essential to keep conducting fundamental research,” says Terzis, a civil engineer at EPFL’s Soil Mechanics Laboratory. His company Medusoil produces organic binders that are similar to biocement.

For the study published recently in Scientific Reports, Terzis worked with scientists from the University of Applied Sciences and Arts of Southern Switzerland to analyze 50 bacteria strains sourced from farmland in Ticino canton. This land is used for grazing dairy cattle and has shown to be particularly well suited for the production of Medusoil’s biocement due to the widely available presence of calcium.

Biocementation relies on stimulating a natural process: The secretion by microorganisms of an enzyme that triggers the formation of carbonate, which then binds with the calcium largely present in the soil to form calcite, a natural cement.

The study identified which naturally occurring strains fabricate the enzyme required for carbonate formation and can be fermented—two factors that make them prime candidates for biocement production. The scientists created a culture of the most promising strain, which was inoculated in a 1.5-meter-high column of sand.

After 24 hours of infiltration, the column was strong enough to sustain its weight and to be used in a variety of geotechnical engineering and geoenvironmental applications, like erosion. The scientists also found that using this strain could cut production costs by 40%.

A paradigm shift

Medusoil, founded seven years ago, supplies organic binders whose carbon impact is at least 55% lower than that of standard cement, which is made by heating an 80% limestone/20% clay mixture to high temperatures. Biocement can be used in a number of geotechnical and building applications, such as to reinforce dams, prevent soil erosion by wind and help protect areas subject to landslides, earthquakes or cyclic loads induced by road and railway traffic.

To test yet another application, the company’s biocement was used in a project in Geneva to recover concrete aggregates from demolished buildings. And because biocement can be employed several times, it supports the circular economy.

In the Scientific Reports study, the authors note that this naturally occurring biocementation process can be applied on a large scale and can help drive a paradigm shift towards greater sustainability in the construction industry.

New production plant

Medusoil reached a new milestone in 2024 with the opening of a production plant in Molondin, near Yverdon-les-Bains. “The plant can generate 400,000 liters of biocement per year, which is enough to stabilize five kilometers of riverbank against erosion,” says Vincent Laurençon, Medusoil’s head of manufacturing.

The company also has a mobile biocementation plant designed to make use of local raw materials. It was recently transported by truck to Romania, for example, where it was employed to reinforce roads. The firm intends to pursue its cutting-edge R&D and has projects lined up this year in France, the Middle East and the Netherlands.

Decoding emotions in seven hoofed species with AI

Can artificial intelligence help us understand what animals feel? A pioneering study suggests the answer is yes. Researchers from the Department of Biology at the University of Copenhagen have successfully trained a machine-learning model to distinguish between positive and negative emotions in seven different ungulate species, including cows, pigs, and wild boars. By analyzing the acoustic patterns of their vocalizations, the model achieved an impressive accuracy of 89.49%, marking the first cross-species study to detect emotional valence using AI.

“This breakthrough provides solid evidence that AI can decode emotions across multiple species based on vocal patterns. It has the potential to revolutionize animal welfare, livestock management, and conservation, allowing us to monitor animals’ emotions in real time,” says Élodie F. Briefer, Associate Professor at the Department of Biology and last author of the study.

The work is published in the journal iScience.

AI as a universal animal emotion translator

By analyzing thousands of vocalizations from ungulates in different emotional states, the researchers identified key acoustic indicators of emotional valence. The most important predictors of whether an emotion was positive or negative included changes in duration, energy distribution, fundamental frequency, and amplitude modulation. Remarkably, these patterns were somewhat consistent across species, suggesting that fundamental vocal expressions of emotions are evolutionarily conserved.

The study’s findings have far-reaching implications. The AI-powered classification model could be used to develop automated tools for real-time monitoring of animal emotions, transforming the way we approach livestock management, veterinary care, and conservation efforts. Briefer explains, “Understanding how animals express emotions can help us improve their well-being. If we can detect stress or discomfort early, we can intervene before it escalates. Equally important, we could also promote positive emotions. This would be a game-changer for animal welfare.”

Key findings include:

  • High accuracy—The AI model classified emotional valence with an overall accuracy of 89.49%, demonstrating its strong ability to distinguish between positive and negative states.
  • Universal acoustic patterns—Key predictors of emotional valence were consistent across species, indicating an evolutionarily conserved emotional expression system.
  • New perspectives on emotional communication—This research offers insights into the evolutionary origins of human language and could reshape our understanding of animal emotions.

To support further studies, the researchers have made their database of labeled emotional calls from the seven ungulate species publicly available.

“We want this to be a resource for other scientists. By making the data open access, we hope to accelerate research into how AI can help us better understand animals and improve their welfare,” Briefer concludes.

This study brings us one step closer to a future where technology allows us to understand and respond to animal emotions—offering exciting new possibilities for science, animal welfare, and conservation.

Roche Announces SBX Technology, Creates Sequencing Buzz

A few years ago, a new NGS platform being announced at the Advances in Genome Biology and Technology (AGBT) meeting would not have been a surprise; multiple new instruments entered the NGS arena over the span of a few years. In turn, the genomics community has grown used to teams touting their “game-changing” platforms, with promises of either lower cost, longer reads, higher accuracy, more throughput (or all of the above!) But those days had settled down. Or so we thought.

This week, Roche made it feel like 2022 again. Just days before the AGBT meeting kicks off in Marco Island, FL, the company hosted a much-anticipated webinar to offer a technical introduction to their sequencing expansion technology (SBX): a coming together of two companies that Roche had previously acquired—Stratos Genomics and Genia Technologies.

Why now for the technology unveiling? “It just felt like this was the right timing based on where we were, and our comfort with where we are with the technology which we’re very excited about,” noted Mark Kokoris, vice president, head of SBX technology RS.

Kokoris, who co-founded Stratos Genomics in 2007 and served as the company’s CEO until the acquisition, invented SBX technology. The webinar was designed to be “a heavy dive in on the technical,” he told GEN, with less focus on the timeline of the instrument launch and more focus on “the technology, what we’re putting forward, and what we’re seeing.”

The technology

“Our approach to efficiently sequencing DNA is to not sequence DNA,” Kokoris told GEN, chuckling. Instead, he explained, they created and innovated a biochemical conversion process to change DNA into an expanded surrogate molecule with the idea to rescale the signal-to-noise problem. He brought the idea to his friend Bob McRuer, Stratos’ former CTO, almost 20 years ago, who loved it because it simplified measurement. Then, Kokoris had to figure out how to make the chemistry come together without a roadmap.

Nothing existed, he remembers, to explain how to build a cleavable X-NTP (one of the keys to the technology). When you see the structure of an X-NTP, Kokoris asserted, you probably think, this is crazy! “And I knew it was going to require innovation in protein engineering, molecular engineering, and developing chemistries that didn’t exist. And there was a bunch of other stuff that we didn’t realize at the time that we had to innovate.”

SBX creates a surrogate molecule called an Xpandomer (which is 50 times longer than target DNA) and encodes the DNA sequence information in large, high signal-to-noise reporters. The backbone of the Xpandomer is X-NTPs, which are linked along a target DNA template. The DNA sequence is represented in the X-NTP sequence. The four X-NTP types have a tether that is linked between the base and the alpha phosphate. In the process, the DNA template is degraded and the backbone expands, becoming an Xpandomer which is pulled through a nanopore. This is performed in millions of wells on a CMOS-based sensor.

“When we got to the other side of [building the technology] it was exactly what we had thought. We had solved that single molecule signal to noise. And now that we have transitioned that to a large, 8 million array. It’s bonkers.”

Not Roche’s first rodeo  

Roche is not a newcomer to the sequencing space: the company bought Jonathan Rothberg’s 454 in 2007 for $140 million. 454 had just announced the completion of Jim Watson’s genome (in May 2007) and published in Nature the following the year. But it was evident that 454’s platform could not get the WGS price down significantly, while competition from Illumina and others was rapidly intensifying (PacBio launched at AGBT in 2008). The writing was on the wall when Nature published three back-to-back landmark NGS papers in November 2008—the first African genome, the first Asian genome, and the first cancer genome—all featuring the Illumina platform.

Roche shut down the 454 program in 2013. Meanwhile, Roche had taken a shine to Genia Technologies, the developer of a single molecule, semiconductor-based, DNA sequencing platform using nanopore technology. Roche eventually bought Genia for $350M in 2014.

A crazy idea 

At this point, a new NGS technology has to offer something new or different to users to get noticed. What are the advantages of SBX? Is it read length, cost, accuracy? According to Kokoris, the cost, scale, throughput, and accuracy have all been considered. (The data from the webinar suggested that read accuracy and speed are strengths of the system.) But the one thing that was always the focus—even back when McRuer and he would meet at Starbucks on First Avenue (in Seattle) for six-hour sessions back in 2007—was flexible operation. That was at the top of mind then and “hasn’t changed one bit in 18 years,” noted Kokoris.

“We envisioned a platform that can sequence up and down the throughput spectrum with impunity,” Kokoris explained. “In other words, one system where someone could do four minutes, 40 minutes, or four hours depending on your throughput. That’s what I wanted on one system. And that is why we did single molecule.”

The technology will be released as research use only, but Kokoris noted that there are future ambitions to take it to the clinical setting. This year will be early access, with the plan to commercialize in 2026. No other details on pricing or sample prep were offered at this time.

At the end of the webinar, Kokoris thanked his team and many others. And he thanked those who believed in this when “it seemed like a crazy idea.” Although Kokoris was referring to the technology, it might also seem crazy to some to launch a new NGS technology into an already extremely crowded market. Time will tell. But the SBX announcement has succeeded in creating a buzz that the NGS community hasn’t experienced for a few years. And in the words of Peter Diamandis, MD: “The day before something is a breakthrough, it’s a crazy idea.”

Fragile X Phenotypes Reversed in Mice by Targeting NMDA Receptors

A new study suggests a potential molecular strategy for treating fragile X syndrome, an inherited neurodevelopmental disorder that causes autism spectrum disorder and intellectual disability. This work shows that enhancing the function of the GluN2b subunit of the N-methyl-D-aspartate (NMDA) receptor signaling pathway can correct key neural dysfunctions in a mouse model of fragile X syndrome.

The research, titled “Non-ionotropic signaling through the NMDA receptor GluN2B carboxy-terminal domain drives dendritic spine plasticity and reverses fragile X phenotypes,” was published in Cell Reports.

Led by Mark Bear, PhD, at MIT’s Picower Institute for Learning and Memory, the study builds on previous work by this group exploring the role of NMDA receptors in regulating synaptic plasticity. The group studies synaptic plasticity and has a history of exploring the molecular basis of fragile X syndrome, and a related disorder, tuberous sclerosis (Tsc).

“We ended up here by accident at the end of the last century, following up on studies of the basic neurobiology of synaptic plasticity. One of our discoveries suggested the possibility that we could reverse aspects of fragile X syndrome, and we have stuck with it for 25 years,” Bear shared with GEN.

Fragile X is characterized by excessive protein synthesis, which leads to synaptic dysfunction and predisposition to seizures, while Tsc involves reduced protein synthesis. In fact, crossbreeding mouse models of both conditions results in healthy offspring, with the protein expression levels balancing each other.

NMDA receptors play a critical role in synaptic plasticity, where calcium ions flow through the receptor and contribute to long-term depression. More recent work from the group identified a signaling pathway for NMDA that was independent of ion flow.

The team hypothesized that two subunits of NMDA, GluN2A and GluN2B, have separate functions contributing to the two functional pathways, with GluN2A contributing to synaptic function via ion flow, while GluN2B modifies protein synthesis through the non-ionotropic mechanism.

To test these hypotheses, the team used the shrinkage and enlargement of dendritic spines as a physical marker for synaptic plasticity in response to modifications in NMDA function. Knockout experiments showed that loss of either GluN2A or GluN2B disrupted long-term depression, a consequence of ionotropic signaling. Knocking out GluN2B eliminated spine shrinkage, a hallmark of synaptic plasticity through non-ionotropic signaling.

Bear commented that by using this method, the team “discovered a novel approach to rebalance altered protein synthesis regulation in these diseases, one that had not been known before.”

The MIT researchers genetically engineered mice in which the carboxy-terminal domain (CTD) of GluN2B was swapped with that of GluN2A. They found that disrupting GluN2B’s CTD eliminated its ability to regulate spine size and increased bulk protein synthesis, mirroring what is seen in fragile X syndrome. Conversely, enhancing GluN2B signaling reduced protein synthesis to normal levels, similar to what is observed in Tsc models.

They then tested whether increasing GluN2B signaling could counteract fragile X-like phenotypes. Fragile X model mice were treated with Glyx-13, which selectively binds to GluN2B. Treatment normalized protein synthesis and reduced seizure susceptibility in fragile X model mice, suggesting that targeting GluN2B could represent a novel therapeutic approach for fragile X syndrome.

“These findings suggest that non-ionotropic NMDAR signaling through GluN2B may represent a novel therapeutic target for the treatment of fragile X and related causes of intellectual disability and autism,” the authors wrote.

By putting together data collected over decades, the MIT researchers identified a deeper understanding of a signaling pathway with broad implications for neurological disorders. Bear pointed out to GEN that “fragile X and TSC are monogenic causes of autism spectrum disorder.” He continued, “We expect that insights gained by studying these diseases will be broadly applicable.”

“One of the things I find most satisfying about this study is that the pieces of the puzzle fit so nicely into what had come before,” said Bear.

AI system predicts protein fragments that can bind to or inhibit a target

All biological function is dependent on how different proteins interact with each other. Protein-protein interactions facilitate everything from transcribing DNA and controlling cell division to higher-level functions in complex organisms.

Much remains unclear, however, about how these functions are orchestrated on the molecular level, and how proteins interact with each other—either with other proteins or with copies of themselves.

Recent findings have revealed that small protein fragments have a lot of functional potential. Even though they are incomplete pieces, short stretches of amino acids can still bind to interfaces of a target protein, recapitulating native interactions. Through this process, they can alter that protein’s function or disrupt its interactions with other proteins.

Protein fragments could therefore empower both basic research on protein interactions and cellular processes, and could potentially have therapeutic applications.

Recently published in Proceedings of the National Academy of Sciences, a new method developed in the Massachusetts Institute of Technology Department of Biology builds on existing artificial intelligence models to computationally predict protein fragments that can bind to and inhibit full-length proteins in E. coli. Theoretically, this tool could lead to genetically encodable inhibitors against any protein.

The work was done in the lab of associate professor of biology and Howard Hughes Medical Institute investigator Gene-Wei Li in collaboration with the lab of Jay A. Stein (1968) Professor of Biology, professor of biological engineering, and department head Amy Keating.

Leveraging machine learning

The program, called FragFold, leverages AlphaFold, an AI model that has led to phenomenal advancements in biology in recent years due to its ability to predict protein folding and protein interactions.

The goal of the project was to predict fragment inhibitors, which is a novel application of AlphaFold. The researchers on this project confirmed experimentally that more than half of FragFold’s predictions for binding or inhibition were accurate, even when researchers had no previous structural data on the mechanisms of those interactions.

“Our results suggest that this is a generalizable approach to find binding modes that are likely to inhibit protein function, including for novel protein targets, and you can use these predictions as a starting point for further experiments,” says co-first and corresponding author Andrew Savinov, a postdoc in the Li Lab. “We can really apply this to proteins without known functions, without known interactions, without even known structures, and we can put some credence in these models we’re developing.”

One example is FtsZ, a protein that is key for cell division. It is well-studied but contains a region that is intrinsically disordered, and therefore, especially challenging to study. Disordered proteins are dynamic, and their functional interactions are very likely fleeting—occurring so briefly that current structural biology tools can’t capture a single structure or interaction.

The researchers leveraged FragFold to explore the activity of fragments of FtsZ, including fragments of the intrinsically disordered region, to identify several new binding interactions with various proteins. This leap in understanding confirms and expands upon previous experiments measuring FtsZ’s biological activity.

This progress is significant in part because it was made without solving the disordered region’s structure, and because it exhibits the potential power of FragFold.

“This is one example of how AlphaFold is fundamentally changing how we can study molecular and cell biology,” Keating says. “Creative applications of AI methods, such as our work on FragFold, open up unexpected capabilities and new research directions.”

Inhibition, and beyond

The researchers accomplished these predictions by computationally fragmenting each protein and then modeling how those fragments would bind to interaction partners they thought were relevant.

They compared the maps of predicted binding across the entire sequence to the effects of those same fragments in living cells, determined using high-throughput experimental measurements in which millions of cells each produce one type of protein fragment.

AlphaFold uses co-evolutionary information to predict folding, and typically evaluates the evolutionary history of proteins using something called multiple sequence alignments (MSAs) for every single prediction run. The MSAs are critical, but are a bottleneck for large-scale predictions—they can take a prohibitive amount of time and computational power.

For FragFold, the researchers instead pre-calculated the MSA for a full-length protein once, and used that result to guide the predictions for each fragment of that full-length protein.

Savinov, together with Keating Lab alumnus Sebastian Swanson, Ph.D., predicted inhibitory fragments of a diverse set of proteins in addition to FtsZ. Among the interactions they explored was a complex between lipopolysaccharide transport proteins LptF and LptG. A protein fragment of LptG inhibited this interaction, presumably disrupting the delivery of lipopolysaccharide, which is a crucial component of the E. coli outer cell membrane essential for cellular fitness.

“The big surprise was that we can predict binding with such high accuracy and, in fact, often predict binding that corresponds to inhibition,” Savinov says. “For every protein we’ve looked at, we’ve been able to find inhibitors.”

The researchers initially focused on protein fragments as inhibitors because whether a fragment could block an essential function in cells is a relatively simple outcome to measure systematically. Looking forward, Savinov is also interested in exploring fragment function outside inhibition, such as fragments that can stabilize the protein they bind to, enhance or alter its function, or trigger protein degradation.

Design, in principle

This research is a starting point for developing a systemic understanding of cellular design principles, and what elements deep-learning models may be drawing on to make accurate predictions.

“There’s a broader, further-reaching goal that we’re building towards,” Savinov says. “Now that we can predict them, can we use the data we have from predictions and experiments to pull out the salient features to figure out what AlphaFold has actually learned about what makes a good inhibitor?”

Savinov and collaborators also delved further into how protein fragments bind, exploring other protein interactions and mutating specific residues to see how those interactions change how the fragment interacts with its target.

Experimentally examining the behavior of thousands of mutated fragments within cells, an approach known as deep mutational scanning, revealed key amino acids that are responsible for inhibition. In some cases, the mutated fragments were even more potent inhibitors than their natural, full-length sequences.

“Unlike previous methods, we are not limited to identifying fragments in experimental structural data,” says Swanson. “The core strength of this work is the interplay between high-throughput experimental inhibition data and the predicted structural models: the experimental data guide us towards the fragments that are particularly interesting, while the structural models predicted by FragFold provide a specific, testable hypothesis for how the fragments function on a molecular level.”

Savinov is excited about the future of this approach and its myriad applications.

“By creating compact, genetically encodable binders, FragFold opens a wide range of possibilities to manipulate protein function,” Li agrees. “We can imagine delivering functionalized fragments that can modify native proteins, change their subcellular localization, and even reprogram them to create new tools for studying cell biology and treating diseases.”

How mosquitoes hear may inspire new ways to detect natural disasters

One of nature’s most disliked creatures may very well unlock a breakthrough in disaster response. A multidisciplinary Purdue University research team is recreating mosquito antennae to better study their sensitivity to vibrations. Should the research prove fruitful, it could lead to improvements in monitoring and detecting natural disasters such as earthquakes and tsunamis.

Research groups under Purdue professors Pablo Zavattieri and Ximena Bernal conducted this work, which is published in the journal Acta Biomaterialia.

“We’re still in the early stages but we’re pretty optimistic that we’ll at least learn a great deal,” said Zavattieri, the Jerry M. and Lynda T. Engelhardt Professor of Civil Engineering in Purdue’s College of Engineering. “Taking inspiration from nature and using it to advance scientific research has been a core feature of engineering since the very beginning.”

Despite lacking traditional ears, mosquitoes rely on their antennae to navigate the auditory landscape, homing in on crucial sounds amid the background noise of their own wingbeats.

Through analysis of the mosquitoes’ antennal features—particularly the arrangement and morphology of sensory hairs—civil and construction engineering Ph.D. candidate and team researcher Phani Saketh Dasika said they have already gained profound insights into how these adaptations enhance the auditory sensitivity and selective response to environmental cues.

“Using advanced micro-CT imaging to create high-fidelity CAD models for finite element analysis, we found that the architectural features of mosquito antennae enable species- and sex-specific acoustic target detection, even amid nontarget signals like their own wingbeats,” Dasika said. “Our findings also suggest that mosquito antennae are capable of detecting a broader range of frequencies than previously thought, though not all of these may be actively utilized.”

The team’s findings have provided key information for determining whether a mosquito’s antennae could inform the design of acoustic sensors.

“By modeling and contrasting the response of the antennae of species of mosquito using sound for different purposes, hearing mates or eavesdropping on frogs, we were able to tease apart features modulating hearing sensitivity and tuning,” said Bernal, a professor of biological sciences in Purdue’s College of Science. “Understanding how these structures work is the first step toward developing acoustic sensors inspired by their sensitive antennae.”

In terms of societal impact, insights from mosquito antennae could also inform the development of smart noise-canceling materials, Zavattieri said. These materials, potentially incorporating microfluidic channels or tunable metamaterials, could be used to create soundproofing panels for buildings, noise-canceling headphones or even acoustic cloaking devices.

Arc Institute’s AI Model Evo 2 Designs the Genetic Code Across All Domains of Life

“Today, we can for all practical purposes read, write, and edit any sequence of DNA, but we cannot compose it. Maybe we can cut and paste pieces from nature’s compositions, but we don’t know how to write the bars for a single enzymatic passage. However, evolution does.” —Frances Arnold, PhD (Nobel Prize Lecture 2018)

Evo, the genome foundation model developed by the Arc Institute published last November that generalizes across the languages of biology — DNA, RNA, and proteins for both predictive and generative capabilities — has received a major update.

In a new preprint that is not yet peer-reviewed and first published on Arc’s website, Evo 2 moves beyond single-cell genomes of bacteria and archaea to include information from humans, plants, and other more complex single-celled and multi-cellular species in the eukaryotic domain of life.

The model’s resulting research applications span a diverse array of scientific fields including drug discovery, agriculture, industrial biotechnology, and material science. The multimodal and multiscale work is a collaboration with Nvidia along with contributors from Stanford University, UC Berkeley, and UC San Francisco.

“The recipe for life is entirely present in the genetic information contained in our DNA,” said Kimberly Powell, vice president of healthcare at Nvidia. “We’re seeking a deeper understanding of biological complexity. Evolution has solved this problem over millions of years, and Evo 2 aims to learn from this knowledge.”

In healthcare, understanding which gene variants are tied to a disease is an invaluable tool for therapeutics. Early validation of Evo 2’s capabilities showed that the model can identify how genetic mutations affect protein, RNA, and organismal fitness. In tests with variants of BRCA1, a gene associated with breast and ovarian cancer risk, Evo 2 achieved greater than 90% accuracy in predicting which mutations are benign versus disease-causing. 

Patrick Hsu, PhD, Arc Institute co-founder and an assistant professor of bioengineering at UC Berkeley, stated that Evo 2 is the only model that can predict the effects of both coding and noncoding mutations.

“It is the second-best model for coding mutations, but it is state-of-the-art for noncoding mutations, which other variant effect prediction methods, such as AlphaMissense from DeepMind, cannot score,” said Hsu. 

Hsu also described Evo 1 as a “blurry picture of single-cell life” because it was trained on a corpus of 300 billion nucleotides derived from prokaryotic genomes. The team “wanted to be much more ambitious” in this collaboration with Nvidia. 

Evo 2 was built on NVIDIA’s DGX Cloud platform and is trained on more than 9.3 trillion nucleotides from the genomes of more than 128,000 species across the tree of life. The model uses a novel architecture called StripedHyena 2, which enabled training that was “nearly three times faster than optimized transformer models,” according to Dave Burke, PhD, chief technology officer at Arc Institute. The model also has 40 billion parameters and is similar in scale to the current generation of large language models released from Meta, DeepMind, or OpenAI. 

Evo 2 can process DNA sequences of up to 1 million nucleotides at once, allowing it to understand relationships between distant parts of the genome. Hsu stated that this long context length unlocks multiple molecular scales, from short biological molecules, such as tRNA, or clusters of genes (e.g., operons), to entire bacterial genomes or eukaryotic chromosomes. 

Arc Institute and Nvidia describe Evo 2 as the “largest publicly available AI model for biology to date.” Evo 2 is available for public use on the NVIDIA BioNeMo platform and as an interactive user-friendly interface called Evo Designer. In addition, the authors have made its training data, training and inference code, and model weights open source. 
Biology’s app store Understanding biology as a “language” is not a new concept. Advances in genome sequencing have allowed us to “read” the human genome, while the invention of CRISPR technology expanded our toolbox to gene “editing.”  

In 2023, Hsu and Brian Hie, PhD, assistant professor of chemical engineering at Stanford University, began thinking about designing or “writing” biological sequences, including proteins, by starting at the foundational layer of DNA itself. “After all, proteins themselves are encoded directly by the genome,” emphasized Hsu.  

“Machine learning started to revolutionize biology, and models such as AlphaFold or ESMFold enabled protein structure prediction and design. Despite these advances, the complexity of these molecules is dwarfed by the overall complexity of an entire cell,” Hsu continued. 

Given that biological functions are not accomplished by a single protein molecule in isolation, constructing synthetic genomes can provide a valuable research tool to investigate broader biological context, a feat that Evo 2 is tackling head-on. 

“A lot of biological design until now has focused on the molecular level because that’s all that we could control. If we have a powerful model that lets us generate at the scale of complete organisms, then that unlocks a lot of downstream tasks [with a wide array of use cases],” said Hie. 

The Evo 2 preprint described three design tasks that span different levels of genomic complexity: 1) mitochondrial genome 2) prokaryotic genome of Mycoplasma genitalium, a commonly used model of the minimum genome, and 3) yeast chromosome, which represents eukaryotic organisms.

For all three design tasks, the preprint showed evidence supporting genome coherence, such as the construction of genes that code for all the components of the electron transport chain (as predicted by AlphaFold 3) in the case of the mitochondrial genome, and the presence of natural homologs and more complex genomic architecture, such as introns, in the case of the yeast chromosome. 

The preprint also presented a workflow for “generative epigenomics,” which designed DNA sequences with desirable chromatin accessibility profiles to simulate eukaryotic gene regulation.  

When asked about plans for experimental validation, Hie stated that a collaboration with large DNA synthesis and assembly experts from the University of Washington is underway to insert the chromatin accessibility designs into mouse cells for validation studies. 

Looking ahead, the Arc Institute is interested in building on this biological complexity by constructing the virtual cell.  

“The bottleneck to drug discovery is that we don’t know what causes the disease to begin with,” said Hie. “If we have a very capable model of the genome and we couple this with information from the environment through RNA sequencing, gene regulatory networks, and cell signaling networks, then this combined multimodal framework will let us answer these fundamental questions about disease.”

Hie sees Evo 2 as an “operating system”, or a foundational layer, that provides a platform for broad generative functional genomics. While Evo 2 “might not solve all questions in biology,” the model offers a wider breadth of applicability compared to task-specific predecessors, such as AlphaFold for protein structure prediction. 

“We want to empower the research community to build on top of these foundation models. That’s why we put in so much effort with Nvidia to make this fully open source,” weighed in Hsu. “We’re really looking forward to how scientists and engineers build on this ‘app store’ for biology.” 

error: Content is protected !!