Researchers from the Wellcome Sanger Institute, Cambridge University Hospitals NHS Foundation Trust (CUH), and Open Targets have created a detailed cellular study of Crohn’s disease (CD), mapping how gene activity changes across more than 50 cell types in the gut. (Founded in 2014, Open Targets is a pre-competitive, public-private partnership that uses human genetics and genomics data to systematically identify and prioritize drug targets.)
Co-led by Tim Raine, MD, PhD, consultant gastroenterologist at Cambridge University Hospitals NHS Foundation Trust, the team analyzed over a million gut cells from people with Crohn’s and from healthy controls, comparing changes in the gut lining and identifying immune cells that drive inflammation. The resulting single cell RNA-sequencing (sc-RNA-seq) resource, IBDverse, characterizes each cell type and those whose activity shifts in disease, uncovering new molecular and cellular signatures of immune activity in the gut lining.
Co-first author Monika Krzak, PhD, formerly at the Wellcome Sanger Institute and now based at the Institute of Metabolic Science, University of Cambridge, said, “Crohn’s disease is complex, variable and deeply personal to every individual living with it, which is why understanding it at the level of single cells is so important. By creating this unprecedented map of more than one million gut cells, we are giving researchers around the world a powerful new tool to uncover how inflammation begins, persists and may one day be stopped. This is the kind of open science that can accelerate discoveries and bring us closer to better treatments for patients.”
Raine added, “There is an urgent need for increased understanding of the biology of Crohn’s disease if we are to develop more effective and safe medications for people living with this condition. The patients who contributed to this research have helped us build insight into the different ways that gut function and immune function are disrupted in the disease, and with the insight comes immediate new avenues for drug development and targeted therapies.”
In their newly published report the team concluded, “These findings establish a comprehensive cellular and molecular framework for CD, offering insights into disease mechanisms and therapeutic opportunities.”
Inflammatory bowel disease (IBD) is an umbrella term used to describe disorders that cause chronic inflammation of the gastrointestinal tract. Over half a million people in the U.K. are estimated to be living with IBD, which includes Crohn’s disease and ulcerative colitis.
Crohn’s is a chronic condition that causes inflammation and ulcers in the digestive tract, from the mouth to the anus, often affecting the small intestine and colon. However, the authors noted, “Although inflammation is most commonly observed in the terminal ileum, CD exhibits substantial heterogeneity in disease location, severity and behavior, both between patients and within patients, over time.”
Although inflammation is most commonly observed in the terminal ileum—the final section of the small intestine—Crohn’s is found in many locations of the body and with variation in severity both between patients and within patients over time.
While therapies targeting immune cells have improved clinical outcomes for some patients, non-response to treatment remains high, with 15% of Crohn’s patients requiring surgery within five years of diagnosis. Consequently, there is an urgent need to better understand the etiology of CD in order to broaden therapeutic opportunities,” the investigators stated.
For their newly reported study the researchers took and analyzed biopsies from 111 patients with Crohn’s and a history of current or previous terminal ileitis—inflammation of the ileum—and 232 healthy volunteers. The team performed single-cell RNA sequencing to measure gene expression in individual cells. “Single-cell RNA sequencing (scRNA-seq) technologies provide a high-throughput means to dissect complex tissues at the resolution of single cells and cell types,” they noted.
By creating a comprehensive map of cellular and molecular differences in Crohn’s compared to healthy controls, the researchers established IBDverse as a result—an online data resource of over 1,185,000 cells isolated from small intestine samples. The large IBDverse dataset will serve as an open resource for future research.
Using the data, the scientists identified genes that are abnormally expressed in Crohn’s and those where expression is specific to given cell types and cellular processes.
One of the study’s most striking findings was a ‘molecular scar’ in the gut lining. Even after visible inflammation had healed, genes that help send messages to the immune system stayed switched on in the gut’s stem cells—the cells that constantly renew the lining. This suggests that an episode of inflammation leaves a lasting mark on these cells, which may shape how the gut responds to inflammation in future. In their paper the authors noted, “We uncovered epithelial changes marked by interferon-driven upregulation of major histocompatibility complex class I molecules that persisted in progenitor cells after macroscopic inflammation resolution.”
The researchers also identified a population of macrophages—immune cells that engulf and digest cellular debris—with high expression of the gene ITGA4. These cells were key drivers of inflammation through the JAK/STAT pathway, which carries signals from the cell surface to the nucleus to switch genes on and off. “ITGA4+ macrophages were identified as key inflammatory drivers, showing enriched JAK–STAT signaling and cytokine expression (interleukin-6 (IL-6), IL-12 and IL-23),” the investigators stated. Drugs that block this pathway, known as JAK inhibitors, are already used to treat IBD, which points to these macrophages as a likely target of therapies.
Co-first author Tobi Alegbe, PhD, at the Wellcome Sanger Institute and Open Targets, said “For inflammatory bowel diseases like Crohn’s and ulcerative colitis, it’s still unclear what is going wrong in the gut cells to cause inflammation. We have been able to compare gut cells of hundreds of people with and without IBD. This has given us new insight into the genes and cell types that are involved during active disease, and lays the groundwork for similar approaches to understand diseases of other major organs like eczema and asthma.”
Co-senior author Carl Anderson, PhD, at the Wellcome Sanger Institute, said “What makes this study different is that we designed replication in from the start and found that even with hundreds of patients and standardized protocols, fewer than half of the gene expression changes we detected in one cohort replicated in the other. That’s a sobering finding for the field. The biology that did replicate consistently points to the gut lining itself as a key player in Crohn’s with a molecular signature in epithelial cells that persists even after inflammation has resolved. We don’t yet know what that persistence means, but it likely influences how the gut responds to future inflammatory insults.”
European biotechs presented a range of interesting technologies and novel therapeutic biotechs at the recent BioTrinity conference, organized by BioUK, formerly OBN. The meeting highlighted a spectrum of cell and protein-based therapies, as well as novel delivery methods, with eye-catching investments and partnering opportunities.
One of the stand-out technologies discussed by U.K.-based University of Southampton spin-out Renovos Biologics was the use of synthetic nanoclays. However, these are not being used for their traditional application of controlling drug release, but rather, for medical delivery. According to Agnieszka Janeczek, PhD, “We are developing an injectable, biodegradable nanoclay for use with bone morphogenetic protein 2 (BMP-2), a crucial growth factor in bone regeneration. We intend to use this on patients who have had spinal fusions after trauma or sports injuries.”
Current treatments to encourage bone growth after spinal fusion, such as BMP-2 in its current formulations, are poorly retained around the spine, causing inflammation or, worse, bone growth outside the spine. This often leads to patients needing revision surgery at a cost of up to $50,000. “BMP-2 was used around fusions in the cervical spine, but the growth of bone and inflammation caused some patients to develop breathing difficulties and has resulted in a black box warning against using BMP-2 for this application,” noted Janeczek.
To overcome these issues, the company has developed a synthetic nanoclay that provides a localized environment conducive to cell infiltration. The nanoclay retains the BMP-2 bioactive molecule until newly regenerated tissue gradually replaces it.
“We have used RENOVITE® with BMP-2 so that this protein is injectable with a 23-gauge needle to deliver BMP-2 in a gel to allow precise templating of new bone formation, and therefore safer and more efficient bone fusion,” she continued. “Using our nanoclay, we see better quality bone growth as the bone grows evenly around and through the fusion, unlike BMP-2 on its own, which can sometimes initiate bone growth around the fusion to create an eggshell effect with bone only on the outside.”
Renovos Biologics, which is developing an injectable, biodegradable nanoclay for use with bone morphogenetic protein 2, intends to use this crucial growth factor in patients who have had spinal fusions after trauma or sports injuries. [Vadym Terelyuk/Getty Images]
The Renovos nanoclay retains the BMP-2 protein more readily around the fusion. As a result, it is possible to administer BMP-2 in a lower dosage, giving it the potential benefit of use in lower-cost markets, said Janeczek, and concluded that “Twenty-one percent of lumbar fusions are in patients that are younger than 45 years old, and the market for bone fusion could be worth $24.5 billion by 2035. Our lead asset RENOVITE BMP-2 will enter first-in-human trials by 2027, and we welcome investment to help assess this game-changing product in the clinic.”
Patching up wounds
Thomas Hafner, CEO of Onya Therapeutics, based in Ebbw Vale, U.K., stated that “For 25 years, wound care has been an innovation desert because therapies are difficult to scale and are not easy to adapt. This is not a glamorous area, but it is a massive opportunity.”
To address this issue, several companies are developing devices, artificial skin, or anti-microbials. Others are developing sponges for managing exuding wounds, making them thicker for better absorption. As Hafner pointed out, they are doing this without asking, “Is this the right approach? “Wound care is a massive problem. For example, 30% of patients with diabetic foot ulcers will lose a limb within five years. Globally, there are triple the number of patients with chronic wounds as there are with cancer—wound care is a $400 billion crisis in plain sight.”
Currently, with wound care, 90% of the costs are for the labor to manage the wound and only 6% for the products themselves. To address this, Onya says it is developing a treatment that could collapse the 90% costs of wound management by accelerating healing. The company calls it active exudate therapy.
“In chronic wounds, the amount of exudate increases, and so instead of mopping up the exudate with dressings, we have found a way of turning off the tap. By doing this, we encourage healing and reduce amputations, potentially saving billions of dollars,” Hafner noted.
Onya, which closed a £2.6 million seed financing round in 2025, is utilizing a compound known as OTX-PP01, based on potassium permanganate, a molecule with long-established clinical safety and broad antimicrobial and astringent action. This compound is delivered via a patch. “Unlike other wound management treatments, we don’t absorb the fluid; instead, our OTX-PP01 patch supports reduction of excessive exudate by addressing chronic inflammation, and creating the conditions for healing to restart,” said Hafner.
Onya Therapeutics’ OTX-PP01 patch supports reduction of excessive exudate by addressing chronic inflammation, and creating the conditions for healing to restart. [Victor Golmer/Getty Images]
The compound does this in two ways: first, it targets exudate with an astringent action that can reduce excessive wound fluid at the source, targeting the inflammatory cycle that can stall healing. Second, it addresses inflammation using a broad-spectrum antimicrobial action targeting pathogens through oxidation.
“OTX-PP01 is not a new molecule. It has over 150 years of safe clinical use, but the delivery method is new,” explained Hafner. “The patch is easy to fit, and a nurse can put it on the wound and leave it on for 15 minutes for the compound to absorb and do its job of healing. After removing the patch, a nurse can then redress the wound.”
Onya has designed its OTX-PP01 patch product to treat diabetic foot ulcers, venous leg ulcers, and pressure ulcers. It does not have to undertake Phase I studies as these were waived based on the established safety profile of the active compound. “We have a Phase II/III adaptive trial design, which we hope to complete in the next four to five years. Our aim with OTX-PP01 is to transform wound management to wound healing,” said Hafner.
Eyes on the prize
Therapies for treating ophthalmic diseases were also noteworthy at BioTrinity, with StemSight and Link Biologics presenting data on their promising clinical candidates. StemSight, based in Tampere, Finland, is a biotech company spun out from Tampere University. The firm is developing off-the-shelf induced pluripotent stem cell (iPSC) therapies to cure corneal blindness.
According to Laura Koivusalo, PhD, CEO and founder of StemSight, the market value of regenerative medicine products across a variety of indications was $63 billion in 2026. However, the bottleneck for making these therapies widely available is manufacturing and scaleup. Koivusalo said that “Recently in Japan, the first iPSC treatments have been approved for use, and the advantage of using iPSCs combined with biomaterials is that they provide efficient and durable therapies.”
Her company is developing a therapy to treat limbal stem cell deficiency (LSCD), a rare disease that causes blurry vision due to corneal epithelium loss and eventually blindness. There are over 240,000 sufferers worldwide, and there are currently no treatments for this condition available for most patients.
“The only treatment option for LSCD is available only to patients with one healthy eye,” Koivusalo explained. “Scientists can harvest limbal stem cells from the healthy eye and transplant them into the diseased one to restore the damaged surface cells. The therapy proves that the cell transplantation approach works, but this is a personalized autologous treatment. Additionally, if both eyes have LSCD, and the patient is truly blind, this treatment is not possible as there are no healthy stem cells available to harvest. That’s where the high value of cure is really measured.”
StemSight is developing an allogenic limbal stem cell therapy, which uses gene-edited iPSCs to reduce immunogenicity. Currently, the firm can produce one hundred patient doses per batch using a GMP-compliant process. It then freezes these cells until their delivery on a biomaterial carrier for patients.
StemSight and Link Biologics presented data in their talks on their promising clinical candidates for ophthalmic diseases. [Shironosov/Getty Images]
“Our process produces high-quality stem cells for a low cost of goods,” continued Koivusalo. “We are currently in late preclinical development with our lead product, STE-101, and have shown that we can regenerate corneal epithelium in rodent models using imaging and histology. These are objective measures of efficacy, as we cannot ask rodents if they can read a chart.”
StemSight is aiming to treat ten patients in a Phase I/II study starting early 2028.
Unlike StemSight, Link Biologics, a U.K.-based spin-out from the University of Manchester, is developing first-in-class biologics to treat Dry Eye Disease (DED) and Wet Age-related Macular Degeneration (AMD). The firm is developing therapies based on TSG-6 (Tumor Necrosis Factor-α-Stimulated Gene/Protein-6), a secreted glycoprotein that has an endogenous role to protect tissues from inflammatory damage and promote repair. Link’s TSG-6-based therapies have enhanced activities compared to the native protein and thus have a unique combination of anti-inflammatory, tissue-protective, and tissue-reparative properties.
Reuben Dawkins, CEO and co-founder of Link Biologics, stated that “We are using these protein biologics to treat DED and wet AMD because these are billion-dollar markets where there are limited treatment options available. Current standard of care for DED, for example, involves cyclosporine, which has a 15% response rate, or Xiidra, where 50% of pivotal trials did not show improvement in key efficacy measures vs no treatment. Overall, nine in ten DED patients stop using their initial medication within one year.”
The company’s lead candidate, LB001, is in preclinical development for the treatment of DED. Dawkins presented data to show that in mouse models, twice daily administration of LB001 for seven days reduced corneal epithelial damage and suppressed inflammatory markers compared to a branded cyclosporine (Restasis).
“Unlike other treatments, LB001 has a dual tissue-repair and anti-inflammatory action which may make this molecule more effective than current therapies,” noted Dawkins. “We have completed CMC so are able to manufacture LB001 to GMP standards, and it is stable as an eye drop formulation. We will be taking LB001 into a 180-patient Phase I/II clinical trial, which we aim to commence in 2027.”
Sue Pearson, PhD, is a freelance writer based in the U.K.
At the end of a tense and scoreless first half of a soccer match between the English men’s team and rival Germany, millions of Brits let out a collective sigh and did what they so often do in moments of stress: They made tea. That wave of electric kettles clicking on, however, caused a different kind of stress: a huge and sudden increase in demand for electricity. But National Grid, which operates the local transmission network, was ready.
Just as those kettles started heating up, an AI program sent instructions to a data center in London to slow down some of the facility’s power-hungry chips. This reduction helped make sure there was enough supply to match demand, staving off potential blackouts or damage to electrical hardware. For data centers, which normally guzzle power without consideration for anyone or anything else’s needs, it was a radical departure.
It was also a simulation. In December 2025, engineers sought to test a new breed of data center built to be flexible about its electricity needs, so they re-created the energy demand facing the UK’s grid during a match from the 2020 Euro tournament. They wanted to see how their software, called Conductor, would have responded had it been online at the time.
Conductor is the signature product of Emerald AI, a firm based in Washington, DC, that’s part of a wave of companies trying to figure out whether data centers can work within the confines of the existing electric grid.
This year, Emerald is set to deploy Conductor in a new facility in the part of Virginia known as Data Center Alley, this time connected to the live grid. When overall demand spikes, Conductor will turn down the power used by the data center, while making sure its servers still carry out their timeliest and most important jobs. Emerald’s partners on the project—which include Nvidia and the giant data-center operator Digital Realty—bill it as one of the world’s first “power-flexible AI factories.”
Demonstrating that data centers can participate in this kind of give-and-take could ease what many tech leaders identify as the bottleneck in getting facilities online: It takes far longer to get approval for, construct, and connect new power plants than to build data centers. PJM, the grid operator in Virginia and the largest one in the US, for instance, needs eight years to bring new generation online, according to RMI, an energy research and advocacy group. “We need to solve the energy equation,” says Josh Parker, head of sustainability at Nvidia. “AI factory flexibility is the bridge between the incredible demand for AI and the immediate limitations of our energy grid.”
Speed, though, is only one of the issues. Once facilities do plug in, neighbors often criticize them for drawing too much electricity and contributing to rising prices. They say the data centers generate more noise than they do long-term jobs, contribute to pollution, and threaten to put people out of work. Organizers stalled over $150 billion worth of projects in 2025, according to Data Center Watch, and policymakers alert to the public mood are starting to impose limitations on development.
More than a dozen states are considering bans, and local moratoriums are in effect in places like Minneapolis and DeKalb County in Georgia. At the federal level, the GRID Act, a bipartisan bill in the US Senate, proposes to sever new data centers from public grids entirely. Some operators are already moving that way by trying to develop their own power generation.
Rather than rushing to build new power plants, companies could find part of the solution to the crunch right under our noses—or, more precisely, in the transmission lines under our feet and above our heads. The existing system operates near its full capacity during only a small number of high-demand hours throughout the year. This means, some grid experts argue, that if data centers can limit the power they draw during those stretches, they won’t need to wait for big infrastructure upgrades or build their own off-grid generation.
Indeed, a growing number of studies have shown there could be plenty of power available for data centers that can flex. A widely discussed 2025 report from researchers at Duke University found that the US grid could offer an additional 76 gigawatts—about 5% of its entire capacity, and about enough to accommodate projected data-center growth in the US through 2030—to facilities that are willing to reduce their usage just 0.25% of the time. That’s about 22 hours a year. And when researchers from Princeton University and two grid-modernization companies looked at locations for new data centers in the PJM region, their report, which was funded by Google, found that a 500-megawatt facility capable of flexing for less than 1% of the year could reach full operation three to five years faster than one that’s inflexible.
Flexible power connections could also help data centers address some of their PR problems. By decreasing their draw at times of grid stress, for instance, they could avoid diverting power from where it’s most needed, thus boosting stability. By using existing capacity, they might be able to reduce the need for new fossil-fuel power plants and spread fixed costs over more electricity users, pushing prices down.
The AI power pinch is attracting resources and research into strategies for grid flexibility overall, which could help negotiate a tricky period: Taken together with electric vehicles, air-conditioning, and other sectors, data centers are helping drive what analysts predict will be a 25% increase in US electricity demand by 2030 compared with 2023 levels.
Ideally, flexibility gives grid operators more control over the flow of electrons, making them leaders of a harmonious ensemble rather than hostages to inflexible electricity requirements. That will help them manage demand spikes across the entire system and deal more effectively with the intermittent nature of renewables like wind and solar. “Demand flexibility is incredibly useful for power grids,” says Johanna Mathieu, a grid expert at the University of Michigan. “It helps reduce electricity costs and improve grid reliability.”
But while advocates see plenty of benefits, the concept brings complexity. For data centers, compromising on energy needs can be a hard sell. Flexibility requires utilities and grid operators, which tend to be operationally conservative, to change long-held practices. And some skeptics also say that flexibility distracts from the very real need to build more grid infrastructure faster, and could even pose risks to our electricity supply.
Still, some technologists, grid operators, and utilities are hoping to show that flexibility works—not only in white papers or simulations but in real life.
The poster children for data-center growth default toward inflexibility. Hyperscalers like Microsoft and Oracle have proposed enormous new centers, many of which would rely on off-grid, natural-gas-burning power plants. When xAI wanted to speed up the buildout of the Colossus site outside Memphis, Tennessee, it rolled up with gas turbines on flatbed trucks. The facility, now in operation, is facing blowback from regulators and residents about the spike it’s causing in emissions and other pollution. In any case, there aren’t enough gas turbines worldwide to meet the demand from data-center operators.
One big obstacle for anyone demanding a lot of power is that our grids are mostly rigid. They’re designed to supply enough power to meet total demand when it’s highest, even if that’s for only a relatively small number of hours a year. That conservative approach is a simple route to reliability, but it means that the grid has quite a bit of headroom. “The grid is already overbuilt by a lot. If you were an airline running at 30% utilization, you would not buy more planes,” says Amit Narayan, the cofounder and CEO of GridCare, a company developing flexibility technologies, referring to a 2025 Stanford study of transmission lines in western North America. “If you are running a grid at 30% utilization, there’s no scientific reason you can’t go to 60.”
“If you were an airline running at 30% utilization, you would not buy more planes. If you are running a grid at 30% utilization, there’s no scientific reason you can’t go to 60.”
To be fair, the idea of flexibility isn’t entirely foreign to grid operators. For decades, they’ve practiced a technique called demand response: When it looks as if demand will get too close to supply, as it might during a heat wave when many people turn on the AC at the same time, they call large commercial or industrial facilities and ask them to shut down parts of their operations. This method can help avoid the need to fire up so-called peaker plants, which run on fossil fuels, but it’s slow, imprecise, and hard to scale.
In the 2000s, as the adoption of technologies like electric cars and solar panels presented new challenges, more internet-connected grids also provided new means of flexibility. Virtual power plants, or VPPs, offered a smarter, faster, more granular alternative. Electricity customers ranging from factories to homeowners with smart thermostats, solar panels, or big batteries would allow the utility to adjust their draw to help meet demand—often getting paid for their (frequently unnoticed) trouble.
After the generative AI boom began with the release of ChatGPT in 2022, some companies began to see flexibility as a way to get data centers set up more easily, efficiently, and affordably. If they bring AI money into existing grids and reduce or defer the need for expensive upgrades, data centers could actually help spread out fixed costs so as to lower rates for other users. A study from Duke University published this past February, for instance, found that flexibility could reduce rates by 0.5% to 2.8%.
PETRA PÉTERFFY
The trick is figuring out how data centers, notorious power hogs, can keep operating when their flexible connections are throttled. Flexibility specialists envision three possible ways. The simplest is for the new data center to install on-site backup power storage or generation to tap when the grid is maxed out—at their own expense, of course.
A facility could also fill the gap by drawing on a VPP. The utility would turn down the electricity going to users who signed up for the VPP, and the data center would pay them for their flexibility. This method wouldn’t require any major infrastructure, but it would require the utility to have a big VPP program and to coordinate the exchange at a time when the grid was under stress. While VPPs exist to some extent in nearly 40 states, the rules governing them vary widely, and they are empowered to do more in some areas than in others.
Finally, a data center could simply use less power at peak times. The conventional wisdom is that they won’t go for such limits, particularly when every number-crunching server can feel like a goose potentially laying little golden eggs. But some experts are betting that the value of getting up and running quickly is enough to change their minds. “There is a clear and growing trend,” says Ayse Coskun, chief scientist at Emerald AI. “Operators are increasingly willing to trade some level of flexibility for faster grid interconnection.”
GridCare, a startup based in Silicon Valley, was one of the first companies to use flexibility to get data centers online quickly. Instead of looking at grids only in worst-case scenarios when electricity demand is highest, the company analyzes the system under all conditions, explains CEO Narayan, who studied smart grids at Stanford. It feeds every part of the grid—including power plants, lines, substations, and homes—into a generative AI model that creates a “digital twin” for different grid configurations. It then picks out results that could unlock capacity while maintaining reliability, and it feeds those into another model trained on the physics of electrical components like resistors and capacitors to make sure they’re realistic.
GridCare found its first customer in the Silicon Forest, an area in the Pacific Northwest named for the trees that dominate the landscape and the IT industry that has more recently sprouted up there. The local grid needed more capacity to support more data centers. “Data centers wanted ‘speed to power,’” says Isaac Barrow, a manager of data-center relations at Portland General Electric, or PGE, the local power generator and distributor, “but transmission buildout is a long process that’s very costly.”
In 2024, Aligned Data Centers came to PGE wanting to expand its operation in Hillsboro, Oregon, and PGE followed a recommendation from GridCare. Aligned will install a 31-megawatt battery, set to be in service in May 2027, and decrease its draw by up to that amount when the grid becomes congested. Bundled with other flexibility measures, that battery has allowed PGE to increase the capacity it can offer Aligned and other nearby operators by 80 megawatts without any new power plants. Though the buildout of data centers in Hillsboro has faced plenty of pushback from locals, Barrow points out that it could have the knock-on effect of lowering costs for ratepayers, because it spreads out the tab.
Other companies are promoting different flavors of flexibility. Google has been moving processing loads from facilities in areas experiencing demand spikes to those in less stressed spots since 2023. It’s signed agreements with five utilities, including the Tennessee Valley Authority and Indiana Michigan Power, that add as much as a gigawatt of flexibility.
Voltus, a major VPP provider across the US and Canada, markets a “bring your own capacity” program in which a data-center company can fund a VPP nearby. The grid operator can use the VPP to decrease demand at busy times, and participants get a financial thank-you. “We can spin up new VPPs on the order of months,” says Emily Orvis, Voltus’s vice president of energy markets. In June, the company signed their first such data-center deal: a three-year plan in which Google will bankroll a VPP in the PJM interconnection.
Of all the approaches to flexibility, Emerald AI’s may be the most ambitious: asking data centers to dial into the grid’s needs. The company’s Conductor software, which can run on premises or in the cloud, builds on the research of chief scientist Coskun. Her group at Boston University showed in a pair of 2013 papers that a data center could watch the grid and help balance big power fluctuations, such as the intermittent effects of solar and wind power. By 2022, she and her colleagues had tested their methods on a cluster of 36 research servers and shown that the system could respect power limits without breaking the processes it was running.
One of the most important questions for Conductor is deciding which AI processes can be slowed down to save energy without kneecapping performance. A lot of companies label their jobs by priority—a real-time chatbot query, for instance, might outrank something like a web search that’s part of a deep research project. When they don’t, Emerald AI tries to infer priority from the nature of the job. Conductor then analyzes the AI workload to determine how tweaking the power to a given processor will affect the performance and help meet the usage limits set by the grid operator.
“The performance curve changes for different kinds of workloads,” says Coskun. “Each AI job is going to have a different location on that curve. Our intelligence is figuring out where you are on that curve.”
PETRA PÉTERFFY
Last year, Emerald AI began assessing the technology’s readiness for real-world use in a series of tests, raising the difficulty each time. The trials were carried out in partnership with the Data Center Flexible Load Initiative—a collaboration among tech companies like Google and Nvidia, utilities like Duke Energy, and grid operators like PJM that aims to help establish a repeatable framework for power-flexible data centers.
The first challenge was in Phoenix, a fast-growing computing hub. For the test, Conductor took control of a group of server racks laden with 256 Nvidia A100 GPUs—hardware that can use about as much power as around 170 US homes. When presented with a simulation of a busy grid, Conductor reduced the power to the chips by 25% for three hours, while maintaining acceptable computing performance. Emerald AI and its partners reported the results in a paper in Nature Energy in December 2025.
The next trial forced the system to juggle surprise grid fluctuations without advance warning and redirect AI jobs from a data center in Virginia to a less busy one in Chicago. Then, in London, Conductor took the reins of equipment beyond the main GPU processors and faced a more complicated mix of fluctuations, including very short and long bouts of congestion—plus the notorious teakettle effect.
The progress so far shows that flexibility can work, at least in some situations, but only a small fraction of operators have pursued it as yet. “We’re just in the beginning innings of the game,” says Jesse Jenkins, one of the authors of the 2025 Princeton study and cofounder of Firma, a startup that works on data-center flexibility. “People are recognizing that this is a potential solution. The motivation is there; there are some bespoke examples. But there’s no uniform solution set that’s the default option, which is where we need to get.”
While data centers are going up across the US, no place on Earth comes close to the accumulated computing muscle in Northern Virginia’s Data Center Alley. The region is home to around 500 compute-crunching facilities, which represent 13% of the entire world’s capacity; the next two hot spots, Beijing and Oregon, contain 6% each.
There are proposals to build hundreds more facilities in Virginia, but a government study found that the state’s electricity demand will increase 183% (around 26 gigawatts) by 2040 if they all go forward, and supporting even half would be difficult. The power-flexible data center that Emerald AI, Nvidia, Digital Realty, and their partners are building in the suburb of Manassas could demonstrate how data centers can squeeze the power they need out of existing capacity. The facility, slated to come online later this year, is intended to give Conductor the chance to manage power at the largest scale yet and to respond to conditions on a live grid for the first time. In the UK demonstration, Conductor managed a 130-kilowatt AI cluster; in Manassas, it will pull the strings of a 96-megawatt hyperscale AI factory.
Some degree of flex will play a key role as we transition away from fossil fuels and toward a future that has to juggle technologies like solar and wind power, batteries, and electric cars.
For PJM, the Manassas facility points to a potential path through the current power crunch. “We think data-center flexibility, in different forms, will be essential for the reliable integration of data-center load over the short to mid term,” says Scott Baker, who manages demand-side markets at PJM.
But not all grid experts are so sanguine. PJM’s market monitor, which oversees the grid operator, says there are no workarounds when it comes to adding capacity. “The notion that large amounts of data-center load can be added without adding new generation is magical thinking,” says Joseph Bowring, an economist and the head of PJM’s market monitor since 1999.
One problem, he says, is that there’s no way to guarantee that a data center will actually take less power when demand is high. That is, absent any legal or regulatory push for flexibility or compliance, the utility won’t be able to step in to help prevent, say, a blackout. Utilities can rely on resources like power plants, but they can’t control or rely on data centers. “They do not want to be fully interruptible,” Bowring says of the facilities.
Stephen Empedocles, an advisor for technology companies, views flexibility as more of a tool than a silver bullet. “These approaches are excellent for improving grid reliability and getting more out of the infrastructure we already have,” he says, “but they are optimization tools.” They’re not substitutes for the “generation, transmission, and distribution expansion that will still be required,” he continues.
Flexibility advocates agree that over the long term, whether or not AI continues to boom, electrification will drive a need for more generation and transmission. Some degree of flex will play a key role in using grid infrastructure better as we transition away from fossil fuels and toward a future that has to juggle technologies like solar and wind power, batteries, and electric cars. A report published by the International Renewable Energy Agency in January 2026 found that grids around the world will need three times as much flexibility in 2030 as they had in 2019—and 10 times as much by 2050—to balance increasing demand with fluctuating supplies of renewable energy.
The challenge of powering AI could provide just the spark we need to do the work of designing and building smarter, more flexible grids, says Coskun. “I think with a crisis like this, there’s no quick solution,” she says. “Sometimes a crisis like this creates an opportunity to do something differently.”
Amos Zeeberg is a freelance science and technology journalist based in Bucharest. He’s developing a book about technology networks, including electric grids.
Research led by the University of Exeter on a large group of people from the UK Biobank shows that carriers of certain genetic mutations are more likely to have an adverse reaction to oral corticosteroid medications.
The researchers found that a variant in the CYP3A4 gene raised osteoporosis risk and a variant in the CTLA4 gene increased the risk for multiple adverse outcomes including stroke and cataracts.
People with these variants were also more likely to have a higher cumulative exposure to corticosteroids over time, suggesting that genetic differences may blunt the clinical benefit of standard steroid doses, so they end up needing more treatment over time and therefore being at even higher risk of adverse events.
“We were also able to show a clear relationship between the dose of steroid and side effects,” said Deniz Turkmen, PhD, a postdoctoral researcher at the University of Exeter, who presented the work at the European Society of Human Genetics conference in Gothenberg this week. “This precise analysis shows the increased risk associated with long-term treatment.”
Oral corticosteroids are an effective and very common medication used to target excessive inflammation. They are widely used to treat conditions such as asthma, rheumatoid arthritis, lupus, and other autoimmune diseases, as well as severe allergies and some acute respiratory infections.
In the U.S., millions of people take oral steroid tablets each year, and roughly one in 100 adults is on them at any given time. However, even short courses can increase the risk of health problems like infections, fractures and blood clots, and long‑term use is linked to osteoporosis, diabetes, weight gain and cataracts, among other adverse effects.
Around one in 10 people are estimated to have side effects to these medications, but it has been difficult to predict who will respond poorly, which triggered Turkmen and colleagues to look into the genetics of how these drugs are metabolized.
In this study, the team evaluated information from almost 38,000 participants from the UK Biobank. These individuals had primary‑care prescriptions for oral corticosteroids for inflammatory and autoimmune conditions.
For each person, Turkmen and colleagues calculated cumulative steroid dose up to the point they had an adverse event or a follow‑up appointment. They also modelled dose–response relationships with outcomes such as osteoporosis, stroke and cataracts. In total, 21 candidate pharmacogenetic variants were assessed.
Median cumulative exposure to corticosteroid in the study was around 1.5 g. Increasing this dose from 1.5 g to 4.0 g was associated with an approximate 50% higher risk of osteoporosis. Variants in the genes CYP3A4 and CTLA4 were associated with higher cumulative dose and increased risk of adverse outcomes, but only in steroid‑treated individuals, consistent with pharmacogenetic effects on steroid metabolism and immune regulation.
Adding polygenic scores to the evaluation including multiple variants for osteoporosis‑related traits significantly improved prediction of steroid‑related osteoporosis beyond age and sex, with particular improvements seen in individuals under the age of 40 at first prescription.
A genome wide association study identified an area on the genome (HLA‑DQA1) where the allele linked to lower cumulative oral corticosteroid dose was not associated with adverse outcomes, hinting at biology that reduces steroid need without obvious toxicity signals.
“It was reassuring that the genetic findings involving CYP3A4 and CTLA4 aligned with their roles in steroid metabolism and immune regulation, but the improvement in prediction of osteoporosis when we incorporated polygenic risk scores data was remarkable, especially in younger patients,” said Turkmen.
“We hope that, in time, greater availability of genetic data at population level will mean that it will be possible to integrate genomics into everyday healthcare and hence into prescribing decisions. That will be a major step on the road to the provision of personalized medicine for all.”
A research team at the Stanford Medicine-based Wu Tsai Neurosciences Institute has developed a blood-based method that can determine the biological age of more than 40 different cell types throughout the body, providing a new way to assess disease risk and longevity years before symptoms emerge. The method, published in Nature Medicine, combines plasma proteomics with machine learning to develop estimates of the biological age of cell types including neuronal, immune, glial, endocrine, epithelial and musculoskeletal systems.
“With this indicator, we can assess the age of an organ today and predict the odds of your getting a disease associated with that organ 10 years later,” said senior author Tony Wyss-Coray, PhD, professor of neurology and neurological sciences and director of the Knight Initiative for Brain Resilience at the Wu Tsai Neurosciences Institute.
Developing this new cellular aging estimate was driven by the understanding that aging is the largest driver of a range of chronic diseases.
“Over 50% of the global disease burden can be attributed to aging,” the researchers wrote noting that while aging increases the risk of neurodegenerative disease, cancer and other chronic illnesses, “aging itself remains poorly understood, and quantifying its biology is therefore an important priority for improving prevention and treatment.”
This new model builds on the lab’s previous work in which the team developed a blood test capable of estimating the biological age of 11 organs and organ systems: the brain, muscle, heart, lung, arteries, liver, kidneys, pancreas, immune system, intestine and fat. In that study, published last July, the investigators analyzed blood samples from 44,498 participants in the UK Biobank and measured nearly 3,000 proteins circulating in the bloodstream. By identifying proteins associated with specific organs and comparing individual protein signatures with age-adjusted population averages, the researchers created algorithms that assigned biological ages to each organ.
Those organ-specific aging profiles revealed that organs age at different rates within the same person. Approximately one-third of participants had at least one organ classified as “extremely aged” or “extremely youthful,” while one in four had multiple organs falling into those categories. The biological age of a person’s organs was strongly associated with future disease risk. For instance, people with an aged heart were more likely to develop atrial fibrillation or heart failure, while aged lungs predicted a greater risk of chronic obstructive pulmonary disease. An aged brain was linked to a substantially higher risk of Alzheimer’s disease.
“The brain is the gatekeeper of longevity,” said Wyss-Coray. “If you’ve got an old brain, you have an increased likelihood of mortality. If you’ve got a young brain, you’re probably going to live longer.”
Participants with extremely aged brains experienced a 182% increase in mortality risk over roughly 15 years of follow-up in the study, while those with extremely youthful brains had a 40% reduction in mortality risk.
This new research goes beyond organ-level measurements to evaluate aging at a cellular level. For this research, the team analyzed more than 7,000 plasma proteins from 60,542 people across three independent cohorts. Using protein signatures mapped to their likely cellular origins, they constructed machine-learning models capable of estimating the biological age of more than 40 cell types.
The findings showed substantial variation in cellular aging patterns. Between 20% and 25% of people showed accelerated aging in a single cell type, while 1% to 3% showed accelerated aging in 10 or more cell types. These cellular aging signatures were associated with disease status and predicted disease development and mortality over a 15-year follow-up period.
One important finding centered on the risk of developing Alzheimer’s disease. The data showed that people carrying two copies of the APOE4 genetic variant, a major risk factor for Alzheimer’s disease, tended to have older astrocytes in the brain. Extreme astrocyte aging tripled the risk of developing Alzheimer’s disease among APOE4 homozygotes, whereas youthful astrocytes appeared to reduce that risk. “Youthful astrocytes seem to powerfully attenuate the detrimental effects of APOE4,” the researchers wrote.
The researchers also found strong links between cellular aging and amyotrophic lateral sclerosis (ALS). Individuals with extremely aged skeletal muscle cells had a 12.7-fold higher risk of developing ALS than those with youthful muscle-cell profiles. According to the researchers, this elevated risk could be detected more than three years before a clinical diagnosis.
Other findings suggest potential applications in cancer prevention and metabolic disease monitoring. The researchers reported that extreme aging of respiratory epithelial cells identified smokers at higher risk of lung cancer, while aging of myeloid immune cells was associated with elevated risk of developing type 2 diabetes.
“Our findings suggest that specific cellular vulnerabilities and cumulative aging burden influence survival,” the researchers wrote. Skeletal muscle-cell aging showed the strongest association with mortality, followed by neurons, fibroblasts, alveolar type 2 lung cells and myeloid lineage immune cells.
The investigators believe these new findings could begin to shift medicine toward earlier intervention. “This is, ideally, the future of medicine,” Wyss-Coray said. “Today, you go to the doctor because something aches, and they take a look to see what’s broken. We’re trying to shift from sick care to health care and intervene before people get organ-specific disease.”
Kristen Dahlgren never expected to leave her successful 25-year career as a correspondent for NBC News, let alone become part of the fabric advancing cancer medicines. The transition was not driven by professional restlessness but by something far more personal: Dahlgren was diagnosed with stage II breast cancer in 2019.
Kristen Dahlgren Chief External Affairs Officer Parker Institute for Cancer Immunotherapy (PICI)
“I thought I was a lifer at NBC,” Dahlgren told Inside Precision Medicine on leaving NBC News in 2024 to advocate for accelerating the development of cancer vaccines. “I loved my job. But a cancer diagnosis changes you.” That diagnosis would eventually lead her to found the Cancer Vaccine Coalition (CVC).
Her story, however, is only half of a larger narrative. The other half belongs to Parker Institute for Cancer Immunotherapy (PICI) chief executive officer Karen Knudsen, PhD, an oncology leader and healthcare executive whose career has been defined by a single, persistent challenge: how to turn scientific discovery into real-world patient impact.
With the advent of new scientific discoveries, more investment, and increasing public and private support, cancer vaccines are entering a new phase of development, which includes this integration. The goal of these vaccines is to teach the immune system to identify and attack tumors.
Together, Dahlgren and Knudsen aim to accelerate the development and delivery of cancer vaccines, one of the most promising frontiers in cancer research. Their partnership reflects a broader shift in oncology, one that is less about isolated breakthroughs and more about building systems that can deliver those breakthroughs to patients faster.
Two paths, one problem
Before joining PICI, Knudsen had navigated the upper echelons of academic medicine and healthcare leadership. As an oncology healthcare executive for Jefferson Health, Knudsen ran one of the largest health systems in the country. She also led one of the major cancer centers in the United States, a National Cancer Institute (NCI)-designated cancer center, in a role that placed her at the intersection of science and care delivery.
Karen Knudsen, PhD CEO, Parker Institute for Cancer Immunotherapy (PICI)
The experience exposed Knudsen to a stark reality: the greatest challenge in cancer care is not always discovery—it can be cancer care delivery. “I love science, and we are funding more of it,” Knudsen told Inside Precision Medicine. “However, I don’t think that we have a science problem in the U.S. There’s a ton of science. What we have is a translation problem.” Dahlgren arrived at the same conclusion by asking researchers what was preventing them from moving forward and finding that they often felt like they were operating in silos.
What neither woman accepted was the idea that this gap was inevitable. A systemic gap exists between research and patient care, and it is not subtle. Knudsen saw it daily in the clinic. “If you’ve ever run an oncology unit, the first thing you do … is walk the infusion center. The sense of urgency gets much higher.”
Even without new scientific breakthroughs, the system has room for improvement. “Current data indicate that we could reduce cancer mortality by 20% if everyone in the U.S. with a cancer diagnosis received guideline-concordant care,” Knudsen said. “Forget even new discoveries.”
As a patient, Dahlgren encountered a different version of the same problem: she wasn’t informed of cancer vaccines or any related clinical trials. “I did a lot of medical reporting at NBC and had just been treated at one of the top hospitals, but nobody ever mentioned [cancer vaccines] as a possibility,” said Dahlgren. “There just weren’t the clinical trials available to me. I think there is a lack of awareness.”
Dahlgren continued, “The CVC started with a conversation with some doctors who were working on cancer vaccines because, frankly, I didn’t really believe that they were real or that far along because I hadn’t heard anything about it. I think that was part of the problem. There just isn’t a large public awareness of some of the research that’s going on. Even some oncologists are unable to keep up with all the cutting-edge developments.”
Between those two perspectives lies a critical insight: innovation alone is not enough. It must be accessible, scalable, and visible. That insight is at the core of PICI’s model, originally conceived by entrepreneur Sean Parker, co-founder of Napster and the first president of Facebook. With a $250-million grant from The Parker Foundation, PICI is an attempt to redesign the entire cancer immunotherapy pipeline, from discovery to commercialization.
Unlike traditional funding bodies, PICI does not simply award grants but rather establishes research sites across the country, of which there are currently seven. “The agreements allow our investigators, irrespective of geography, to freely share data and materials, so they truly are functioning [as] one institute, multiple geographies,” said Knudsen. “The whole concept here is to advance those studies from preclinical to clinical, obtaining a clinical signal so that other investors and biopharma follow suit toward the goal of near-term patient benefit.”
Cancer vaccines, now!
When Knudsen met Dahlgren at a Milken Institute event, the alignment was immediate. Dahlgren, for her part, had already recognized that PICI was building something she could not replicate alone. “They were already doing everything I wanted to do: building a system where the ecosystem could thrive and where these discoveries could reach patients.”
That meeting led PICI to strategically integrate the CVC to accelerate the development of next-generation cancer vaccines, including personalized neoantigen therapies. “Cancer vaccines have been around, but we really are at this tipping point,” Dahlgren said, who is now at a new role within the PICI as chief external affairs officer, placing vaccines at the top of PICI’s strategic priorities. Knudsen said, “As part of the scientific priorities, vaccines rise right to the top. Cancer vaccines offer a genuine opportunity to revolutionize cancer therapies.”
Several forces have converged to create this moment: advances in genomics, improved understanding of immune biology, and new platforms for vaccine delivery. “We understand the immune system better,” Dahlgren said. “This is the time to fully support this technology and get it through trials, approved, and to patients.” Dahlgren points to emerging data from clinical studies on vaccines across multiple cancer types, she said, “You hear about the work in pancreatic cancer or glioblastoma … patients who have achieved unexpected success.”
For Knudsen, the cancer vaccines won’t just be abundant and accessible; they’ll be personalized. Early clinical results, while small, are compelling, especially for hard-to-treat solid cancers (BOX 1). Knudsen highlighted a 2025 Nature study on kidney cancer led by Catherine Wu, MD, and Toni K. Choueiri, MD. “There were zero recurrences,” said Knudsen, “and everybody in that study got the vaccine. It’s a small study with single-digit patients; I believe there were nine patients. The science is getting there.”
The implications are profound, but so are the challenges. “But then who’s going to take it to the next step in a larger clinical trial?” Knudsen asked. “Biopharma wants the clinical signal.”
That gap between early success and large-scale validation is precisely where PICI aims to operate.
Patients are ready
Part of the problem is financial. The traditional venture ecosystem is not always aligned with early-stage scientific risk. “The venture world is currently locked up due to a lack of IPOs. … They do not have dry powder available to invest capital in opportunities,” Knudsen said.
PICI’s model attempts to bypass that constraint. “We ourselves are a major investor,” Knudsen said. “We’ve put about $400 million into 17 portfolio companies. … Or we just start new companies.” Revenue generated from successful commercialization is reinvested into new research, creating a self-sustaining cycle. “That’s how the nonprofit model works.”
If there is uncertainty in the system, it does not come from patients. “We’re getting reach-outs from patients … ‘What do I need to start to think about so that I can get enrolled in a trial to prevent recurrence?’” Knudsen said. “Who wouldn’t want that?” Dahlgren echoed the urgency from personal experience. “My current course of treatment is to ‘call us when you get a headache,’ and that’s a horrible way to live.”
Cancer vaccines offer a different paradigm, one that emphasizes long-term immune protection. “Imagine if you could get a vaccine … that boosts your T-cell response. … It’s just a matter of taking care of it for the rest of your life.”
Despite growing momentum, cancer vaccines remain misunderstood. “There’s definitely a misunderstanding with the term ‘vaccine,’” Dahlgren said. Knudsen is less concerned about semantics, asserting, “I don’t know if we have a name problem. I don’t think there’s anyone who is going to say, ‘I don’t want to take a vaccine’ when facing cancer.”
Where education does matter, however, is prevention. Knudsen points to the Gardasil vaccine as a case study. “When you reframe that appropriately, as a cancer vaccine … you say actually, we have cancer vaccines that will prevent cancer.” In places like Puerto Rico, Knudsen notes that high uptake of HPV vaccination has already demonstrated what is possible. “They’ll be the first place in the U.S. where cervical cancer will no longer be a thing.”
A shared urgency
While vaccines are a major focus, they are part of a broader transformation in cancer therapy. “Immunotherapy is the same as cancer therapy,” Knudsen said. “Name a cancer modality that does not have a root in the immune system.” This shift indicates a deeper understanding of cancer as an immune-mediated disease, paving the way for more personalized treatments.
What ultimately unites Dahlgren and Knudsen is not just a belief in science but a shared sense of urgency. Dahlgren frames it in deeply personal terms. “There’s no evidence of disease—I’m in remission, and I still think about it every day,” said Dahlgren.
With CVC working lockstep with PICI, Dahlgren and Knudsen are working together to remove those barriers by building a system that does not just generate knowledge but delivers it. For Dahlgren, joining PICI is not a departure from her mission but an amplification of it. “I’ve always looked for the quickest path to these better treatments, and there’s just no doubt that under this new structure that we’ll be able to have a lot more impact.” For Knudsen, it is the continuation of a career spent bridging science and care. “It’s the businessperson in me always looking for how we go faster and farther and make it so that patients benefit more quickly.”
Dahlgren and Knudsen are betting on a future where cancer vaccines and the systems that support them transform not just how cancer is treated but how innovation itself happens. If they’re right, the question will be how quickly these breakthroughs can reach those who need them most.
BOX 1. The Rise of Personalized Cancer Vaccine
Personalized cancer vaccines are rapidly emerging as one of the most promising frontiers in oncology. Unlike traditional therapies, these vaccines are designed for each individual patient, using the unique mutations in their tumor to train the immune system to recognize and destroy cancer cells. Powered by advances in mRNA technology and genomic sequencing, this approach has shifted from theory to real clinical momentum in just a few years.
Early trial results are beginning to validate the strategy. In pancreatic cancer, one of the most difficult diseases to treat, a personalized mRNA vaccine developed by teams led by Vinod P. Balachandran, MD, and Benjamin D. Greenbaum, PhD, at Memorial Sloan Kettering Cancer Center triggered strong, lasting immune responses and extended survival in early-stage studies. In melanoma, combining a personalized vaccine with Keytruda reduced the risk of recurrence or death by 44% to 50%, advancing the approach into large Phase III trials.
Trials in lung, breast, and other tumors show that these vaccines are safe, feasible, and capable of generating strong T-cell responses. Increasingly, they are used in combination with checkpoint inhibitors, acting as a primer that enhances broader immune attack.
The pipeline is growing quickly. Dozens of personalized mRNA and DNA vaccines are now in development, with many expected to reach late-stage trials before 2030. Preventive versions, aimed at stopping cancer before it develops, are also beginning to enter human testing.
Investment is accelerating alongside the science. Pharmaceutical companies, biotech firms, and governments are committing billions to personalized vaccine platforms. In the United Kingdom, a partnership between the National Health Service and BioNTech aims to treat up to 10,000 patients with individualized vaccines by 2030.
Challenges remain, including small trial sizes, complex manufacturing, and the need for large-scale validation. Still, the trajectory is clear. Personalized cancer vaccines are moving steadily toward mainstream clinical use.
For decades, cancer medicine has been driven primarily by genetics—the search for mutations, deletions, amplifications, and other DNA changes that directly alter how cells behave. Precision oncology has largely focused on identifying these genomic drivers and matching them with targeted therapies. But another biological layer has steadily moved from the margins of research into the center of clinical care: epigenetics.
Epigenetics refers to chemical modifications that regulate how genes are turned on or off without changing the underlying DNA sequence itself. One of the best-known examples is DNA methylation, in which small chemical tags called methyl groups attach to DNA and influence whether genes are active or silenced. These marks help explain one of biology’s most fundamental questions: how a brain cell and a heart cell can contain the exact same genetic code yet function in completely different ways.
Rather than rewriting DNA, epigenetics controls how a stretch of DNA is read. It shapes normal development, tissue specialization, immune responses, aging, and disease progression. In cancer, epigenetic disruption can silence tumor-suppressor genes, activate harmful pathways, and create stable molecular fingerprints that reveal where a tumor came from and how it behaves. Increasingly, those fingerprints are becoming clinically actionable.
Today, epigenetics is helping physicians classify brain tumors more accurately, detect pancreatic cancer from a simple blood draw, improve molecular sequencing workflows, and even therapeutically silence disease-causing genes. Across diagnostics and therapeutics, the field is rapidly becoming one of the most practical and powerful tools in precision medicine.
Few people have watched that transformation more closely than Matija Snuderl, MD, a neuropathologist at NYU Langone. He remembers a time when epigenetics barely registered in medical education.
Matija Snuderl, MD Neuropathologist, NYU Langone
“When I was in medical school,” he said, “there were literally like two pages on epigenetics in the genetics book, and it had no implications for cancer whatsoever.”
That changed quickly.
Brain tumors become a clinical test case
Snuderl became deeply interested in epigenetics around 2013, when researchers began realizing that epigenetic patterns could do far more than explain developmental biology. They could also reveal a tumor’s cell of origin and, in some cancers, directly illuminate the mechanisms driving carcinogenesis.
Brain tumors proved to be one of the most compelling settings for that work. “Brain tumors are only about two percent of all cancers,” Snuderl explained, “but they are disproportionately heterogeneous.”
That heterogeneity is extraordinary. There are now roughly 180 recognized molecular subtypes of central nervous system tumors. Yet for decades, many of those diseases were classified almost entirely by histology—how the tumor looked under a microscope. Tumors with similar visual features were grouped together, even when they were biologically distinct.
The result was predictable. “We ended up putting a lot of tumors that are completely different into the same basket, treating them the same way, and then we were surprised that we didn’t see results,” Snuderl said.
Epigenetics offered a way out of that problem. Because DNA methylation reflects both developmental lineage and disease biology, it provides a stable molecular fingerprint of tumor identity. Unlike RNA, which fluctuates significantly and degrades quickly after tissue collection, methylation patterns are remarkably durable. “The beauty of methylation,” Snuderl said, “is that it’s an incredibly robust biomarker.”
A molecular fingerprint for diagnosis
Snuderl began a collaboration with the German Cancer Research Center (DKFZ) in Heidelberg and proposed a simple but powerful hypothesis: every tumor type has a distinct methylation signature—a recognizable pattern of hypermethylated and hypomethylated regions that functions like a fingerprint.
Using thousands of tumor samples, the team built a machine learning classifier based on a random forest model. The algorithm was trained to recognize these methylation signatures and assign new tumor samples to the most likely disease subtype. “At every decision point,” Snuderl explained, the classifier essentially asks, “Are you methylated or are you unmethylated?”
The system evaluates more than 10,000 decision trees before generating a calibrated confidence score that indicates how certain the classifier is about a diagnosis.
The visual output is particularly striking. On uniform manifold approximation and projections (UMAPs), which is a method of viewing high-dimensional data in a low-dimensional manner, each tumor appears as a colored dot that clusters with biologically similar tumors. Some groups are extremely tight, reflecting genetically simple and highly consistent diseases. Others are more diffuse, showing greater heterogeneity and multiple biological drivers.
These patterns helped reveal that some long-standing diagnostic categories were fundamentally wrong. One major example involved primitive neuroectodermal tumors (PNETs), once considered a distinct diagnosis. Methylation profiling showed that PNET was largely a “waste basket” category: about 80% of those tumors were misclassified, while the remaining cases represented multiple separate molecular diseases, said Snuderl.
Another example came from low-grade epilepsy-associated tumors in young adults, known as polymorphous low-grade neuroepithelial tumor of the young. It emerged as a clearly distinct clinical and pathological entity through methylation analysis.
These discoveries were not simply academic refinements. They changed how patients were diagnosed and treated.
Taking research to the clinic
Snuderl is quick to point out that many important scientific discoveries never reach patients because researchers stop at publication. A high-impact paper may generate excitement, but translating a method into a clinical diagnostic requires an entirely different level of work—regulatory validation, reproducibility testing, and operational rigor.
“There is very little glory in going through the regulatory framework,” he said. Still, his team decided that it was an essential step.
In 2019, NYU became the first CLIA-certified laboratory in the United States to clinically implement DNA methylation–based brain tumor classification. It was also the first to use machine learning as a primary diagnostic tool for cancer classification.
That achievement was especially significant because New York State has some of the strictest molecular diagnostic regulations in the country. “The benefit of that,” Snuderl said, “is the tests approved in New York State are really rigorous.”
The clinical validation study made clear just how necessary the assay was. Rather than testing only difficult or ambiguous cases, the team prospectively profiled nearly 2,000 consecutive brain tumors to determine real-world utility across all diagnoses.
The results were striking. Approximately 15% of tumors experienced a complete change in diagnosis. Another seven percent were further sub-stratified into clinically meaningful categories. In cases where pathologists could not make a diagnosis at all, methylation profiling resolved the case in more than 85% of patients.
This meant that patients who might have received unnecessary radiation and chemotherapy could avoid overtreatment, while those with aggressive disease who might otherwise have been undertreated were correctly escalated to more appropriate care.
Liquid biopsies and the rise of 5hmC
While brain tumors demonstrate the power of methylation in tissue diagnostics, other companies are applying epigenetics to blood-based cancer detection.
Jeffrey Venstrom, MD CMO, ClearNote Health
At ClearNote Health, Jeffrey Venstrom, MD, CMO, focuses not on traditional 5-methylcytosine (5mC), but on 5-hydroxymethylcytosine (5hmC), a related epigenetic mark that provides a different kind of biological signal.
“While much of the scientific focus on methylation has centered around the 5mC mark,” he said, “we have found the 5hmC mark to be more relevant to the detection of cancer.”
The distinction is important. While 5mC often reflects silenced genes, 5hmC marks genes that are actively being expressed. That makes it particularly useful for identifying active cancer biology, including the upregulated genetic programs associated with tumor growth.
There are also major technical advantages. Because much more of the genome is silenced than active, there is less 5hmC signal to analyze, requiring less sequencing depth to achieve reliable results. Unlike 5mC analysis, 5hmC detection avoids bisulfite conversion, a harsh chemical step that can damage DNA and complicate sequence mapping. Scientific studies have shown that 5hmC is a reliable signal even in early-stage cancers and carries tissue-specific signatures that help identify where a tumor originated.
ClearNote’s platform analyzes cell-free DNA (cfDNA) from a standard blood draw, integrating 5hmC signals with other genomic features and applying machine learning to detect cancer. “Our machine learning model has now been trained on thousands of patient samples and independently validated,” Venstrom said. That platform powers the company’s Avantect® tests for one of the most clinically important targets in oncology: pancreatic cancer.
Pancreatic cancer remains a highly lethal malignancy, largely because it is rarely detected early enough for effective intervention. Standard imaging often misses disease during the narrow window when curative treatment is still possible.
ClearNote designed the Avantect Pancreatic Cancer Test specifically for individuals at elevated risk, including people over age 50 with newly diagnosed type 2 diabetes, those with BRCA mutations, and patients with a strong family history of pancreatic cancer.
The test is a simple blood draw, but its performance is substantial. It now achieves 82.6% sensitivity and 97.5% specificity—metrics that Venstrom says will supplement standard imaging technologies for early detection.
The assay is supported by a robust body of evidence, including multiple conference presentations and peer-reviewed scientific publications, representing more than 20,000 test runs across more than 7,000 patients. It is also being deployed in major global studies, including the U.K.’s National Health Service-led SAFE-D study and the international Pancreatic Cancer Early Detection PRECEDE Consortium.
ClearNote is expanding into ovarian cancer and multi-cancer early detection. Its Avantect Multi-Cancer Detection Test is designed to simultaneously identify eight cancers while predicting tissue of origin. That capability is crucial because, as Snuderl notes of metastatic brain tumors, knowing that cancer is present is only the beginning—knowing where it came from determines how it should be treated. This assay was one of only two selected for the National Cancer Institute’s Vanguard Study, which is evaluating real-world implementation across up to 24,000 participants. For cancers that currently lack practical screening tools, multi-cancer detection might represent the next major leap forward.
Reliable epigenetics in clinics
Paolo Piatti, PhD Director, Zymo Research
From Zymo Research, Xiaojing Yang, PhD, epigenetics group leader, and Paolo Piatti, PhD, director of applied epigenetics, pointed out that one of the biggest barriers to using epigenetics in clinics is ensuring that DNA methylation testing is “reliable, reproducible, and cost-effective,” especially when working with poor-quality samples like cfDNA from plasma, stool, or urine. These samples are often scarce, degraded, and contain only trace amounts of DNA, making it difficult to accurately detect rare epigenetic signals.
The researchers note that bisulfite-based methods continue to outperform newer approaches in difficult clinical settings. Comparative studies, they say, show these methods provide “higher reproducibility and more stable performance” and align better with reference methods such as EPIC arrays. Although newer enzymatic approaches are promising, they might create systematic biases like methylation overestimation. Because of this, they argue that “bisulfite conversion chemistry continues to be reinforced as the clinical gold standard for DNA methylation analysis.”
To address these issues, Zymo developed CE IVD-marked bisulfite conversion kits such as the EZ-96 DNA Methylation Lightning MagPrep. Designed for high-throughput clinical workflows, the kit works with automation platforms like Hamilton, KingFisher Flex, and Tecan Fluent. Its magnetic bead-based system improves DNA recovery and supports downstream PCR, microarrays, and next-generation sequencing.
DNA isolated from a non-small-cell lung cancer was bisulfite-converted using the EZ DNA Methylation-Lightning Kit, then PCR amplified with methylation-specific primers for cancer biomarkers SEPT9, RASSF1A, and Mir129-2, and control primer COL2A1. (L = 50 base-pair marker.) [Zymo Research]
Beyond products, Zymo offers full-service epigenetic workflows like assay optimization, sequencing, and bioinformatics. Their whole-genome bisulfite sequencing service is optimized for low-input and degraded samples like cfDNA and formalin-fixed, paraffin-embedded tissue. These services are supported by “rigorous quality control,” Yang and Piatti noted, and can extend into CLIA/CAP-compliant environments through partner laboratories.
Looking ahead, the company is developing NGS-based workflows that preserve fragmentomic information from cfDNA. This could enable more sensitive tests for “early disease detection and longitudinal monitoring,” Yang and Piatti said, combining both epigenetic and fragmentomic insights for stronger clinical decision-making.
Measuring epigenetics more directly
As epigenetic diagnostics expand, another challenge becomes increasingly important: how to measure these signals accurately and efficiently.Cora Vacher, PhD, director of segment marketing for human genetics at Oxford Nanopore Technologies, argues that traditional workflows often make epigenetics unnecessarily complex.“Oxford Nanopore sequencing can read native DNA and RNA directly, without PCR amplification or chemical conversion,” she said.That means methylation and other epigenetic modifications can be detected alongside the sequence from the same molecule, in the same experiment.
Cora Vacher, PhD Director Oxford Nanopore Technologies
Instead of stitching together multiple assays for sequencing, structural variation, and methylation analysis, researchers can capture all this information simultaneously. “Our technology reads individual DNA or RNA molecules directly as they pass through a nanopore,” Vacher explained, using changes in electrical current to identify both nucleotide sequence and chemical modifications in real time.
Because amplification and chemical conversion are not required, the method preserves long reads and native biological context. It also enables the detection of multiple epigenetic modifications beyond traditional forms of methylation.
In cancer and rare-disease research, this creates a much richer picture of disease biology by linking sequence variation, haplotype phasing, structural variants, and epigenetic regulation on the same molecule. “Because epigenetic information is captured directly,” she said, “the technology is well-suited to developing simpler, more integrated assays that could support future clinical workflows.”
Methylation, represented here by bright areas, appears across the genome and regulates gene expression. Oxford Nanopore sequencers, like the PromethION 2 Integrated, can read these methylated bases directly, alongside typical base information including single point mutations and structural variants.
Therapeutics move beyond diagnosis
Melissa Bonner, PhD CSO, nChroma Bio
Epigenetics is not only improving diagnostics, it is also increasingly becoming the therapy itself. At nChroma Bio, CSO Melissa Bonner, PhD, and CSO Jenny Marlowe, PhD, are developing treatments based on epigenetic silencing rather than gene editing. “Epigenetic silencing enables potent, durable gene silencing without cutting or permanently altering the DNA sequence,” they explained. Instead of editing DNA directly, the approach uses the cell’s own epigenetic machinery—particularly DNA methylation—to turn genes off in a controlled and durable way.
Jenny Marlowe, PhD CDO, nChroma Bio
Their lead program, CRMA-1001, an investigational, clinical-stage therapy, targets chronic hepatitis B, where viral persistence depends on both episomal covalently closed circular DNA (cccDNA) and integrated viral DNA. By permanently methylating both forms, the therapy aims to durably silence viral gene expression and potentially deliver a functional cure.
In preclinical studies, CRMA-1001 produced durable suppression of hepatitis B virus (HBV) biomarkers, with up to 90% of animals achieving complete loss of HBV surface antigen and DNA.
The company is also expanding into cardiometabolic disease, central nervous system, and oncology programs. The shift reflects a larger truth across precision medicine: epigenetics is moving from classification to direct intervention.
Preclinical data suggests CRMA-1001 durably silences hepatitis B virus (HBV) covalently closed circular DNA (cccDNA) and integrated viral DNA at the level of transcription. [nChroma Bio]
The next phase of precision medicine
For Snuderl, the next frontier is clear. After spending more than a decade improving diagnosis, today’s goal is to use that molecular knowledge to improve therapy. “Now we know what we are dealing with and, hopefully, we can find better therapies,” he said.
Future applications of epigenetics might resolve other challenges in today’s healthcare. For example, epigenetics might help identify the origins of metastatic brain tumors when physicians are unable to determine where the cancer began. Epigenetics could also improve clinical-trial enrollment by ensuring that the right patients are matched with the most appropriate studies. “If 15% of patients on your clinical trial have a different type of cancer,” Snuderl said, “there’s a statistically very good chance your trial will fail.”
Precision medicine depends on precision diagnosis. Without accurate classification, targeted therapies and biomarker-driven trials cannot succeed. That is where epigenetics has become indispensable. It provides not just more molecular data, but biological context—revealing lineage, vulnerability, treatment response, and increasingly, therapeutic opportunity. What was once a forgotten footnote in a medical school textbook has become one of the most clinically transformative tools in modern oncology.
Mike May, PhD, is a freelance writer and editor with more than 30 years of experience. He earned an MS in biological engineering from the University of Connecticut and a PhD in neurobiology and behavior from Cornell University. He worked as an associate editor at American Scientist, and he is the author of more than 1,000 articles for clients that include GEN, Nature, Science, Scientific American, and many others. In addition, he served as the editorial director of many publications, including several Nature Outlooks and Scientific American Worldview.
