Google’s medical chatbot can use patient information and photos to pinpoint the cause of a rash. Credit: Adapted from Getty An upgraded version of Google’s medical chatbot can use smartphone photos to diagnose rashes and can evaluate a host of other types of medical imagery — improving the bot’s ability to pinpoint the cause of ailments. An earlier version of the artificial-intelligence (AI) system outperformed physicians in diagnostic accuracy and bedside manner. The upgrade also did better than human doctors at interpreting images such as electrocardiograms and PDFs of laboratory results. The latest version of the system, which is called the Articulate Medical Intelligence Explorer (AMIE) and is still purely experimental, was described in a paper1 posted to the arXiv preprint server 6 May. It has not yet been peer reviewed. The development of systems that integrate images and clinical information “brings us closer to an AI assistant that mirrors how a clinician actually thinks”, says Eleni Linos, director of the Stanford University Center for Digital Health in California, who was not involved in the research. Health-care simulation To test AMIE’s new capabilities, 25 actors playing patients had virtual consultations with the chatbot and a human primary-care physician. The actors simulated 105 medical scenarios involving a range of symptoms and medical histories. They also presented relevant medical images. After each consultation, both AMIE and the human doctor suggested a diagnosis and a treatment plan. A team of 18 specialists in dermatology, cardiology and internal medicine evaluated the performances of the chatbot and the human doctors by analysing transcripts of the visits and the post-visit reports. They found that, in general, AMIE achieved greater accuracy in its diagnoses than did the human doctors. Furthermore, its accuracy was less affected by problems such as low-quality images. Medical school for an AI doctor The updated version of the chatbot is based on Google’s large language model (LLM) Gemini 2.0 Flash, which can process images. The researchers adapted it for medical purposes, by adding an algorithm to improve the LLM’s ability to conduct diagnostic conversations and clinical reasoning. To confirm that the resulting system was indeed superior to the unaltered version of Gemini 2.0, researchers commanded the model to simulate patient–physician dialogues, in which the model plays the parts of a patient, a physician and a third party that evaluates conversations between the other two. “This way, you can sort of imbue it with the right, desirable behaviours when conducting a diagnostic conversation,” says Ryutaro Tanno, a scientist at Google DeepMind in London and a co-author of the study. An AI revolution is brewing in medicine. What will it look like? The method is a departure from earlier versions of the work, which involved retraining the model with specialized databases — a more cumbersome process, says Tanno. “This is much cheaper and potentially [more] accessible.” Scenarios involving actors are also used to evaluate human doctors, says Linos. She notes, however, that simulated interactions can’t capture the complexity of real-world care. “Physicians bring experience, intuition and the ability to physically examine a patient, elements that are hard to replicate in a simulated script.” The results show “promise, but definitely there’s a lot of uncertainty about how it would be helpful”, says Dan Zeltzer, a digital-health specialist at Tel Aviv University in Israel. He also notes that the paper does not detail the codes and prompts used by developers, meaning that other researchers can’t reproduce the system or build on it. Deploying such models in the real world would be a challenge, says Xueyan Mei, an AI scientist at Icahn School of Medicine at Mount Sinai in New York City. “That being said, we do think large language models for diagnosis would be the way to go in the future,” she adds.

Over the past several weeks, hundreds of meeting abstracts have quietly vanished from the websites of research conferences organized by the Lunar and Planetary Institute (LPI) in Houston, Texas — a NASA-funded research institution that maintains key archives and information about US planetary science. The abstracts have a common theme: diversity, equity and inclusion (DEI). These topics were banned from receiving government funding by a 20 January executive order from US President Donald Trump, who called them “illegal and immoral discrimination programs”. The LPI did not seek permission from the abstract authors to remove the material. NASA embraced diversity. Trump’s DEI purge is hitting space scientists hard Many researchers are fuming about the deletions from the archives of a research institute that is managed independently from the federal government. “This is just egregious censorship of science,” says Ingrid Daubar, a planetary scientist at Brown University in Providence, Rhode Island, who co-authored several of the deleted abstracts. Many of the abstracts are pages-long developments of preliminary ideas that are important to the planetary-science literature. Their removals are an “Orwellian rewriting of history”, says a researcher who asked to remain anonymous because of the sensitivity of the situation. Mark Sykes, a senior scientist at the Planetary Science Institute in Tucson, Arizona, called the decision “a profound ethical failure”. Independent organization Several scientists told Nature that they think the LPI and its parent organization, the Universities Space Research Association (USRA) in Washington DC, have over-interpreted the Trump executive order as retroactively applying to independently performed work in public archives. In a statement provided to Nature, the USRA said it “believes that compliance with Administration directives requires DEI content to be out of public view”, unless it is “advised by an authoritative federal body that we can restore that content”. Have Trump’s anti-DEI orders hit private funders? HHMI halts inclusive science programme “Waiting for some ‘authoritative federal body’ to tell them to undo the harm they’ve done is just deflecting accountability,” Daubar says. NASA declined to comment on the basis that it does not run the websites. A group of researchers is now leading a data-rescue effort to locate and rehouse the deleted material on a privately run website. Many federal agencies, including NASA, have removed DEI-related content from their websites in response to the executive order. But the LPI is not a government institution; it was set up in 1968, during the Apollo era, primarily to foster relationships between university scientists and NASA. For decades, it has organized planetary-science conferences, trained researchers and maintained documents and databases on topics such as lunar impact craters. The missing meeting abstracts include research on workforce issues, such as how to build the most effective team of scientists for a mission. One of Daubar’s deleted abstracts described a programme to involve early-career scientists in NASA’s InSight mission to Mars. David Trang, a planetary scientist at the Space Science Institute who is based in Honolulu, Hawaii, had his work on improving mental health among planetary scientists deleted. Removing these abstracts “takes away our culture”, he says. “That’s the saddest part.” The LPI has also scrubbed DEI-related content from the online archives of some of the ‘analysis and assessment’ groups it manages, which provide scientific input to NASA on research topics ranging from Mercury to the outer planets. In some cases, entire reports have been removed, even if DEI was only a small component. Websites removed In January, after Trump issued his executive order, NASA paused meetings of the analysis and assessment groups, and the LPI took the groups’ websites offline. Sites for two of the major groups — specializing in the Moon and in Venus — remain offline, with a notice that they are under review. The others have been partially restored, although sometimes with past reports involving DEI issues deleted. Will US science survive Trump 2.0? Not having all of the analysis-group documents available “is undermining the community’s ability to support NASA’s mission to return to the Moon and travel further to Mars”, says Benjamin Fernando, a lunar geophysicist at Johns Hopkins University in Baltimore, Maryland. The LPI has options besides simply removing content, says Vicky Hamilton, a planetary scientist at the Southwest Research Institute in Boulder, Colorado, who chairs the Mars analysis group. One possibility is adding disclaimers on abstracts or reports with disputed content, as NASA has done in past cases. In its statement, the USRA said it has sought approval from NASA to add such a disclaimer. It might seem odd to worry about conference abstracts when Trump has proposed cutting NASA’s science budget by nearly half, some researchers acknowledge. But many worry about what precedent the abstract removals might set. “It leaves us wondering what’s the next thing that somebody might attack,” says Julie Rathbun, a planetary scientist at Cornell University in Ithaca, New York, who leads the data-rescue effort. “They could say, we don’t like the idea that Mars has water, so now any article that says Mars has water on it, we’re going to take down. It’s all an attack on science.”

LHC experiments don’t create large gold nuggets — but some particles within a beam of lead ions can turn into gold for about a microsecond.Credit: Zen Rial/Getty The dream of seventeenth-century alchemists has been realized by physicists at the Large Hadron Collider (LHC), who have turned lead into gold — albeit for only a fraction of a second and at tremendous cost. The not-so-mysterious transmutation happened at CERN, Europe’s particle-physics laboratory, near Geneva, Switzerland, where the multi-billion-dollar LHC smashes together ions of lead for a portion of each experimental run. How the revamped Large Hadron Collider will hunt for new physics Early chemists hoped to turn abundant lead into precious gold. But differences in proton number between the elements (82 for lead and 79 for gold) made that impossible by chemical means. CERN researchers achieved the feat by aiming beams of lead at each other, travelling at close to the speed of light. The ions occasionally glance past each other, rather than hit head on. When this happens, the intense electromagnetic field around an ion can create a pulse of energy that triggers an oncoming lead nucleus to eject three protons — turning it into gold. The LHC’s ALICE experiment filtered out these instances of transmutation from the wider collision debris. In an analysis published on 7 May in Physical Review Journals1, the team calculated that between 2015 and 2018, collisions at the LHC created 86 billion gold nuclei — around 29 trillionths of a gram. Most of the unstable, fast-moving gold atoms would have lasted around 1 microsecond before smashing into experimental apparatus or breaking into other particles. The ALICE detector at CERN.Credit: Maximillien Brice/CERN Gold is being made any time lead beams are collided at the LHC, but ALICE is the only experiment with the detector set up to spot this process. The analysis “is the first to systematically detect and analyse the signature of gold production at the LHC experimentally”, says Uliana Dmitrieva, a physicist and member of the ALICE collaboration. Another CERN accelerator called the SPS observed lead changing into gold from 2002 to 2004, says Jiangyong Jia, a physicist at Stony Brook University in New York. But the latest experiments are at higher energy, have a much higher probability of creating gold and make for much cleaner observations, he adds. CERN researchers have no plans to take up gold-making as a side hustle, but say that better understanding how photons can change nuclei will help them to improve the LHC’s performance. “Understanding such processes is crucial for controlling beam quality and stability,” says Jia.

The day after Donald Trump moved back into the White House in January, he celebrated a US$500-billion private-sector investment in artificial intelligence (AI) with a high-profile announcement in the Roosevelt Room. The new president looked on as technology billionaire Larry Ellison highlighted one of the initiative’s most transformative goals: using messenger RNA vaccines to transform cancer treatments. By harnessing AI to analyse tumour genetics, Ellison explained, researchers could rapidly design personalized vaccines tailored to an individual’s cancer. “This is the promise of AI and the promise of the future,” he said. Are the Trump team’s actions affecting your research? How to contact Nature Biotechnology executives were elated. Trump had, just five years earlier, propelled mRNA medicines into the spotlight through his signature effort to fast-track the development of a coronavirus vaccine. Now, just one day into his second term, he was once again elevating the technology to the national stage. “Then the bottom fell out,” says Deborah Day Barbara, co-founder of the Alliance for mRNA Medicines (AMM), a trade group representing more than 75 companies and academic institutions that are advancing mRNA research, development and manufacturing. A prominent vaccine critic who had vilified the mRNA-based COVID-19 jabs, Robert F. Kennedy Jr, was appointed to lead the country’s top health agency, and long-time champions of immunization science in the civil-service sector were shown the door. Research grants tied to HIV prevention and pandemic preparedness were abruptly cancelled, including many involving mRNA. And numerous other projects that were focused on mRNA vaccine technology were compiled into a list, potentially signalling their impending termination. At the same time, legislators in several states have been pushing to ban or restrict the use of mRNA-based medicines for infectious diseases. None of these measures has become law, but the efforts threaten to destabilize the mRNA industry, creating uncertainty and potentially limiting patient access to emerging treatments. The anti-mRNA sentiment — coupled with the sweeping shake-up of science funding across the United States — has sparked fears that this once-celebrated technology, widely seen as a major engine of next-generation vaccines and therapeutics, could soon find itself on the chopping block. For AMM executive director Clay Alspach, a principal at Leavitt Partners, a health-care consulting firm in Washington DC, the message has been unmistakable: “This is an existential threat,” he says. By mid-March, the AMM was holding regular conference calls to strategize. Members swapped intelligence, compared notes on delayed grants and tried to anticipate what might come next. Amid the uncertainty, a few questions loomed large: how far would the clampdown on mRNA go? Would it stop at COVID-19 jabs? Would it extend to all vaccines in development for influenza and other infectious threats? Or reach even into mRNA-based drug therapies in the works for cancer, autoimmune disorders, rare genetic diseases and more? From hero to zero Five years ago, the US government was spending billions of dollars to support the development, manufacturing and roll-out of mRNA vaccines, which played a major part in curbing the COVID-19 pandemic. Pharmaceutical companies were pouring in capital and building ambitious pipelines centred on mRNA. The technology was feted with a Nobel prize. Investor confidence was sky-high. Now, in the span of just a few months, the mood across the industry has grown darker — chilled by a newly hostile political climate. Will US science survive Trump 2.0? One contract manufacturer of mRNA products has seen a substantial decline in business as government-backed vaccine programmes have their funding pulled, according to a senior company executive. Another biotech executive says that his mRNA-focused company is considering relocating planned clinical trials for anti-viral vaccines to outside the United States — or scrapping them entirely, shifting the firm’s focus to less politically volatile therapeutic areas. “It’s all just a commercial and regulatory risk now,” he says. Both executives requested anonymity to avoid drawing political attention to their companies — but their experiences reflect a broad upheaval that is now rippling through the industry. In a survey released this month by the AMM, nearly half of 106 senior biotech and pharma executives reported direct impacts from US policy shifts this year — including project downsizing, budget cuts, delayed investments, terminated partnerships, job losses, hiring freezes and planned relocation of operations overseas1 (see ‘An industry at risk’). Source: Ref. 1 Much of the current antipathy towards mRNA vaccines can be traced back to the COVID-19 pandemic, and the political and cultural backlash it left in its wake. Critics cite the compressed timelines and emergency use authorizations as signs that the safety of the vaccines was compromised. Vaccine mandates — imposed by governments, employers and schools — further fuelled resentment. Meanwhile, conspiracy theories about DNA alteration and population control continue to circulate widely online, deepening public mistrust and giving political traction to opposition against mRNA technology. What began as fringe scepticism has increasingly entered the mainstream consciousness, amplified by partisan media and political figures who frame the vaccines not as public-health tools but as symbols of government overreach. Among the most prominent of these is Kennedy, now secretary of the US Department of Health and Human Services (HHS), who has long questioned the safety of childhood immunizations and built his political brand on vaccine resistance. (HHS officials did not respond to requests for comment. A spokesperson for the White House pointed to a public statement that did not address the questions posed by Nature.) Exclusive: NIH to terminate hundreds of active research grants There are some conservative voices who are more supportive of mRNA technology — for example, a February report2 from the Texas Conservative Coalition Research Institute, a public-policy organization in Austin, urges policymakers to recognize the technology’s broad potential in medicine and agriculture. But those views have been mostly drowned out by more extreme rhetoric. Even the term ‘mRNA’ has become a political lightning rod; its charged connotations now influence scientific discourse and health policy far beyond the vaccine debate. “That paranoia has gotten wrapped into mRNA as a word,” says Jeff Coller, an RNA biologist at Johns Hopkins University in Baltimore, Maryland, who is also involved in several small biotech firms. Seeking to reframe the narrative, Coller and others are mobilizing around a strategic communications offensive, emphasizing mRNA’s potential not just in thwarting infectious threats but also in treating many of the same chronic conditions targeted by Kennedy’s ‘Make America Healthy Again’ initiative. The campaign to rehabilitate mRNA’s reputation starts at the top: with appeals to Trump’s legacy as a champion of medical innovation. Legacy on the line AMM leaders are preparing to publish a series of editorials that make the case that Trump’s decisive leadership during Operation Warp Speed — the 2020 programme that delivered COVID-19 vaccines in record time — marked the beginning of a new chapter in biotechnology and positioned the United States at the forefront of what many see as the fourth great wave of pharmaceutical innovation, after small-molecule drugs, biologics and cell and gene therapies. Framing it as a chance for Trump to cement his place in medical history, they are urging the president to build on the foundation that he helped to lay. In particular, they pointed out that, by supporting mRNA-based cancer treatments, he could achieve a major unmet goal that was advanced by his predecessor Joe Biden, who had made “ending cancer as we know it” a signature priority. Trump “could be the president who is a true visionary on cancer”, says Coller, a leading academic voice at the AMM. People queue for COVID-19 vaccinations in Washington DC.Credit: Jim Watson/AFP via Getty Such messaging just might resonate. Although Trump criticized the roll-out and mandates surrounding the COVID-19 vaccines in the period between his two terms, allies say he remains proud of the part he played in accelerating the technology’s development. “President Trump believes that Operation Warp Speed was a roaring success, and that the COVID mRNA vaccines were his great achievements,” says Robert Malone, a scientist involved in foundational mRNA research and a high-profile voice in conservative-leaning health-policy circles. But leading the charge against mRNA technology are individuals in the ‘medical freedom movement’ — Kennedy chief among them. They contend that COVID-19 vaccines were rushed through approval without adequate long-term testing, alleging that safety corners were cut in the name of speed, and that the risks of mRNA platforms continue to be deliberately downplayed. At his confirmation hearing earlier this year, for example, Kennedy — who previously described an mRNA-based COVID-19 jab as the “deadliest vaccine ever made” — persisted in claiming that the vaccine was recommended for young children “without any scientific basis”, despite published clinical-trial evidence3 to the contrary. Public-health researchers and vaccine scientists emphasize that mRNA vaccines have consistently demonstrated robust safety and effectiveness in preventing severe COVID-19 outcomes, supported by extensive data from rigorous clinical trials and real-world studies. Yet, with trust in institutions and in the biomedical establishment crumbling, some argue that pulling back from mRNA is the only responsible course of action — not because the science is flawed, but because the damage to public confidence is too deep. “If mRNA has a chance to have impact in the future, there needs to be a restoration of public trust around it,” says David Mansdoerfer, a political consultant in Fort Worth, Texas, and a former senior HHS official in the first Trump administration. To that end, he, like many associated with Kennedy, would support federal regulators withdrawing approval for all COVID-19 vaccines that initially entered the market under emergency-use provisions — including the mRNA-based ones that later won full approval. Mansdoerfer advocates re-evaluating the jabs under a standard review process. ‘The brand is damaged’ The problem of mRNA’s reputation isn’t just a communications challenge — it’s a systemic liability. “I fear the brand is damaged for most uses,” wrote Vinay Prasad, a haematologist–oncologist at the University of California, San Francisco, in a Substack post in March. A vocal critic of COVID-19 vaccine mandates under Biden, Prasad was selected this month to lead the division of the US Food and Drug Administration that oversees vaccines and other biological products. The branding issue for mRNA is not just a problem in the United States either. An analysis of social-media data across 44 countries, published last year, found “widespread negative sentiment and a global lack of confidence in the safety, effectiveness and trustworthiness of mRNA vaccines and therapeutics”4. mRNA technology helps reinvigorate the hunt for cancer vaccines Nor is the problem confined to the two most prominent mRNA vaccines — those from the biotechnology company Moderna in Cambridge, Massachusetts, and a collaboration between pharmaceutical firms Pfizer in New York City and BioNTech in Mainz, Germany — both developed at record speed in late 2020 and since administered to billions of people worldwide. New concerns have surfaced regarding next-generation jabs, including one from Arcturus Therapeutics in San Diego, California. This ‘self-amplifying’ COVID-19 vaccine, now approved in Japan and Europe, is designed to replicate itself inside cells, allowing for lower doses with fewer side effects. But in Japan, where the vaccine has been available the longest, misinformation has fuelled fears that its components could be transmitted from vaccinated individuals to others, with some rhetoric likening the shot to a ‘third atomic bomb’. Notably, however — unlike in the United States, where the pushback has penetrated policymaking spheres — the leading authorities in Japan have continued to champion the technology. Earlier this year, for example, the Japanese government authorized further domestic manufacturing sites to expand production of the self-amplifying jab — and Joseph Payne, president and chief executive of Arcturus, says there is similar support from other governments around the world. The uniqueness of the US retreat raises an unsettling prospect for the scientists and entrepreneurs who launched the mRNA revolution from US laboratories and start-up companies: the future of the technology might unfold elsewhere. Currently, around two-thirds of the mRNA industry’s jobs are based in the United States, according to the AMM. Yet, if current trends persist, a substantial portion of these positions could be lost or relocated overseas. “I think what we’ll see in the next few years is that this tech is ceded to international competitors,” says Alex Wesselhoeft, who directs RNA therapeutics at the Mass General Brigham Gene and Cell Therapy Institute in Cambridge, Massachusetts. Rewriting the script Aiming to retain global leadership, many researchers in the mRNA sector in the United States are rethinking how to present their innovations — starting with stepping away from the term ‘vaccine’, particularly when describing mRNA-based cancer treatments. Although these therapies work like vaccines, delivering genetic instructions to produce proteins that train the immune system, they are designed to treat disease, not prevent it — hence the industry’s pivot towards labels such as ‘immune therapy’. Ellison, in his January remarks at the White House, still used the legacy terminology of ‘vaccine’ in this context, but companies, including Moderna, are shifting their language. Last month, Moderna — which not only makes one of the world’s leading COVID-19 vaccines but is also a front runner in personalized cancer therapy — updated the product pipeline on its website, replacing ‘cancer vaccines and therapeutics’ with the more neutral label ‘oncology’. NIH has cut one mRNA-vaccine grant. Will more follow? Other companies with diversified mRNA portfolios are also recalibrating their messaging, highlighting opportunities for delivering therapeutic proteins inside the body or components for gene editing to reprogram cells, while downplaying associations with vaccines where possible. Arcturus, for example, is advancing two mRNA-based treatments for rare genetic disorders, with preliminary data from mid-stage clinical trials expected in the coming weeks. According to Payne, “We’ll be speaking more toward our therapeutic programmes, just because those don’t have the same market-based headwinds that our vaccine enterprise does.” “But that doesn’t mean we’re stopping the vaccine efforts at all,” he says. Case in point: earlier this year, Arcturus completed enrolment for a 200-person trial, funded by the US government, to evaluate a self-amplifying mRNA candidate to prevent infection with H5N1 influenza, with results expected later this year. Whether that support will extend beyond the current study, however, remains unclear. Equally unclear — for the entire mRNA sector — are the long-term financial prospects, regulatory pathways and the willingness of governments to stand by a platform they once championed. It doesn’t help that the broader biotech industry is still grappling with a prolonged funding downturn, marked by sluggish capital markets, risk-averse investors and waves of layoffs across the sector. Politics aside, raising money for an early-stage mRNA company to advance promising therapies through clinical trials was already an uphill struggle, notes Ipsita Smolinski, the founder and managing director of Capitol Street, a health-policy consultancy firm in Washington DC. “It’s ugly out there,” she says. Still, many people in the field think the long-term prospects for mRNA technology remain strong — especially as applications in oncology and other areas begin to bear fruit. “This is clearly a rough time,” says vaccine scientist Philip Dormitzer, who helped to develop the mRNA COVID-19 jab at Pfizer and is now based in Canada, where he consults for vaccine makers. “I don’t want to underestimate the real damage that’s been done — it is big.” But between the industry’s coordinated counteroffensive against misinformation and policy threats, and the sustained global momentum behind mRNA innovation, he expects science and public health to prevail. That long view is key, says Payne. “Politics and policies come and go,” he says, “but science always will stand true.”

Lee Zeldin, director of the US Environmental Protection Agency (EPA), announced on 2 May that the EPA would be undergoing a reorganization.Credit: Kayla Bartkowski/Getty The administration of US President Donald Trump has blocked funding for research across the Environmental Protection Agency’s main science division, which employs nearly 1,500 people, according to sources inside the agency and internal e-mails seen by Nature. The messages detail a decision to halt funding for the Office of Research and Development (ORD) and wind down its laboratories, despite a spending agreement enacted by the US Congress in March that funded the EPA at 2024 levels through to September. Democrats on the US House of Representatives Committee on Science, Space, and Technology (CSST) say that the e-mails also contradict statements that EPA representatives provided to them this week. ‘Targeted and belittled’: scientists at US environmental agency speak out as layoffs begin Sent on 7 May, one of the e-mails, which was first reported by E&E News, says that funding for research laboratories has been cut off except for requests related to health and safety. “ORD is shutting down their laboratory activities,” wrote Wayne Cascio, head of the Center for Public Health and Environmental Assessment, an ORD lab in Chapel Hill, North Carolina. “We are unsure if these laboratory activities will continue post-reorganization”, he told his staff members in the e-mail, which he co-signed with the centre’s deputy director Kay Holt. This e-mail follows an announcement by Trump-appointed EPA director Lee Zeldin on 2 May that he plans to reorganize the agency and integrate “scientific staff directly into our program offices” instead of them being siloed in the ORD, “which will better ensure that research directly advances statutory obligations and mission-essential functions”. The ORD carries out research in support of environmental laws and regulations, such as assessing the safety of chemicals in drinking water. The division has long been a target of Trump and his Republican allies who question its science and say it oversteps its authority. EPA representatives met with CSST staff members to discuss the ORD’s reorganization on the same day that Cascio’s e-mail was sent. “EPA explicitly told Science Committee staff … that there wouldn’t be any ‘monumental changes’ to ORD labs,” Zoe Lofgren, a member of the House of Representatives from California and the ranking Democrat on the committee, said in a statement to Nature. “Shutting down ORD laboratories, which are required by law, is as wrongheaded, unlawful, and stupid as it gets,” Lofgren added. Trump proposes unprecedented budget cuts to US science In a statement to Nature, EPA spokespeople said that the leaked e-mail is “factually inaccurate” and that the ORD is not part of the reorganization announced on 2 May. “At ORD and throughout the agency, EPA is continuing research to advance the mission of protecting human health and the environment,” they said. Multiple sources across the ORD, however, told Nature that forward-looking research has already been shut down. A second e-mail seen by Nature, dated 8 May and sent by Timothy Watkins, the ORD’s deputy assistant administrator for management, confirms as much. In the e-mail, which went out to all ORD staff members, Watkins says that all funding requests “should have short time horizons (two to three months) and/or should focus [sic] safety/health and facilities.” “They have basically shut ORD down by cutting off the money,” says Jennifer Orme-Zavaleta, who served as principal deputy assistant administrator at the ORD during Trump’s first presidency, and who is familiar with the situation inside the agency. The internal EPA e-mails come less than a week after the Trump administration released its budget proposal for the fiscal year 2026, which called for a 55% reduction in funding for the EPA and a 46% cut to the ORD. These figures are not final — the US Congress decides how much money to ultimately appropriate. But the news of halted funding at ORD labs plays into fears among some government scientists and science-policy specialists that the Trump team is planning to make big cuts to science spending yet this year rather than wait for Congress to debate the 2026 request. Reapply or resign As part of the 2 May reorganization announcement, the nearly 1,500 people in the ORD — mostly scientists and engineers — have been given until midnight tonight to either apply for a new job within the EPA or take one of two deals that incentivize resignation: an early retirement for those eligible, or a package that provides months of free salary and benefits before termination. Thus far, fewer than 500 new jobs have been posted at the agency, with little to no detail about what the new positions entail. Trump gutted two landmark environmental reports — can researchers save them? Scientists within the agency who spoke to Nature were granted anonymity because of fear of reprisal. They say they suddenly face a ‘Hunger Games’ situation in which, if they stay, they must compete with their colleagues for survival. “I don’t know what I’m going to do,” says one senior scientist in the ORD. They previously intended to stay, but are now debating whether to take an early-retirement deal in hopes that doing so will open up a slot for a younger colleague. “I’m going to be agonizing over this until 11.58 p.m..” Another mid-career scientist who plans to apply for positions in other offices acknowledges feeling frustrated by the fact that some of their senior colleagues are eligible for retirement but are not planning to leave. ORD scientists currently operate independently in their own division. Sources who spoke to Nature argue that it would undermine scientific integrity to move them into other programs under the control of political appointees who are charged with making policy decisions about regulating chemicals, pollution and public health. “This sets the stage for political interference,” says Christopher Frey, an environmental engineer at North Carolina State University in Raleigh who led the ORD under former US president Joe Biden, a Democrat. Trump team orders huge government lay-offs: how science could fare EPA spokespeople told Nature that as of the morning of 8 May, more than 1,800 employees at the agency had applied for one of the two resignation deals; this is on top of the 545 people who took an earlier, government-wide resignation offer three months ago. Prior to Trump taking office, there were roughly 15,000 workers in the agency. ORD scientists fear that anyone remaining in their division who hasn’t found a job elsewhere in the agency will face a mandatory firing called a ‘reduction in force’ (RIF) ordered by Trump’s team across US agencies. One of these scientists worries that their expertise in climate science puts them at a disadvantage, given that Trump has been hostile towards this area of study. “I’m probably not wanted or needed at all,” they say. Despite their long career at the EPA, they aren’t yet eligible for retirement and don’t want to resign. “I would rather make them RIF me, to be honest.”

US researchers who receive consistent grant renewals produce more fresh research than those who don’t, a study finds.Credit: Attila Csaszar/Getty In its bid to reduce government spending and rein in ‘woke’ science, the administration of US President Donald Trump has, over the past three months, terminated about 800 active research grants awarded by the National Institutes of Health (NIH). Now, a study published late last month in the journal Scientometrics1 highlights the long-term benefits of renewing existing grants, finding that US scientists who received renewals from the NIH over the past four decades produced more novel research than those who didn’t. Although the study focused on a major US biomedical-science funding agency, the results might persuade governments worldwide “to implement policies that give stable funding to researchers”, says study co-author Aruhan Bai, a science-policy researcher at the Chinese Academy of Sciences Institutes of Science and Development (CASISD) in Beijing. Continuous funding, Bai says, will ultimately “encourage the real prosperity of science”. Innovation takes time In the new study, Bai and her co-author, Baicun Li, also at CASISD, analysed the scientific output of 642 US scientists who won NIH funding between 1985 and 2021. The study focused on recipients of R01 grants — common, yet highly competitive, awards from the NIH that provide up to five years of financial support and can be renewed several times. As of last year, the success rate of winning an R01 grant was 20%, on average, and the rate for winning a renewal was about 45%. ‘Disruptive’ science has declined — and no one knows why Two subgroups of equal sizes made up the study cohort: one in which researchers received at least a three-year renewal of their first R01 grant without any funding interruption, and a control group of researchers who didn’t receive a renewal and had at least a three-year gap in funding afterwards. To ensure that the two groups were matched in terms of their research quality prior to receiving grants, Bai and Li used the scientists’ career stages, publication histories and measures of their prior research impact to ‘pair’ them. To evaluate the scientists’ output after renewal decisions, the duo used public NIH-funding data, researchers’ publications listed in the biomedical-literature repository PubMed and a newer database called SciSciNet2 that contains scientific publication data and links to related funding and usage information. Bai and Li harnessed these resources to examine the ‘novelty’ of researchers’ publications, calculating a score on the basis of the uniqueness and trendiness of key medical words and concepts in their papers. The pair also measured the ‘switching probability’ for scientists — how likely they were to hop to new research topics — and the ‘diversity’, or spread, of topics they studied using a model that calculated the similarity of their publications over a set period.The results revealed that researchers who received renewals without a funding gap were less likely to switch research topics and investigate a wide array of subjects than were those in the control group, and they produced papers featuring newer or more unconventional combinations of ideas. The team found that renewed funding “helped scientists stay focused” so that they could “dig deeper into one area” of research, Bai says. As a result, she adds, the scientists could generate innovative applications down the line. “Innovation takes more than 10 or 20 years of continuous work,” she says. The importance of continuity Bai and her co-author “did a really nice job” of analysing large databases that provide reliable estimates of scientists’ performance before and after receiving grants, says sociologist Brian Uzzi at Northwestern University in Evanston, Illinois. Still, Uzzi adds, not getting a renewal doesn’t necessarily mean that researchers are not diving deep into their research areas, only that renewals are a tool that gives them more opportunity to do so. At the same time, Uzzi says, shifting research areas frequently could reflect the pressure to chase funding, rather than a path to meaningful scientific outcome. “Before you can actually get to the frontier of an area of knowledge, you’re already forced to switch again,” he adds. Studies like this one demonstrate the importance of sustained science funding, Uzzi says. Not only are the funding cuts at US science agencies probably going to stagnate federal science and lead to brain drain, but failing to renew grants also wastes the money that was already invested, Uzzi adds. If researchers “get their knees cut out from under them” midstream, leaving their work unfinished, Uzzi says, “that’s disruptive to the advancement of science”.

It is a common refrain that we know more about the surface of Mars than the bottom of the oceans. Now, oceanographer Katherine Bell and her collaborators have quantified just how little of the deep seafloor scientists have actually explored: less than 0.001%, an area equivalent to about one-tenth the size of Belgium. The findings were published on 7 May in Science Advances1. The team collected data from about 44,000 deep-sea dives performed either by crewed deep-sea vessels or by remotely operated or autonomous submersibles. Their map shows how dives have focused heavily on waters near a few countries — especially the United States, Japan and New Zealand. Moreover, data from many of those dives are owned by private companies and unavailable to researchers. Vast expanses of the seafloor remain where the only information available comes from rough mapping tools such as sonar, Bell says. “The Indian Ocean is one of the least explored areas.” This map shows the locations of deep-sea dives carried out between 1958 and 2024. Expeditions have focussed on a small number of places, primarily around the United States, Japan and New Zealand.Credit: Ocean Discovery League Bell is the founding director of the Ocean Discovery League in Saunderstown, Rhode Island, a nonprofit organization that is helping to develop affordable remotely operated vehicles and leading an effort to compile a list of 10,000 potential deep-dive sites that could begin to sample a representative portion of the unexplored abyss. Vedachalam Narayanaswamy, an engineering scientist at India’s National Institute of Ocean Technology in Chennai, says the study highlights the need for “more resources to understand the deep-ocean basins”. Narayanaswamy leads the Matsya 6000 project — India’s first crewed deep-sea submersible, which can carry three people down to a depth of 6,000 metres inside a 28-tonne titanium alloy sphere.

Pseudomonas aeruginosa bacteria (blue) is harmful to people and often found in hospitals.Credit: Juergen Berger/Science Photo Library A strain of bacterium that often causes infections in hospital can break down plastic, research published this week in Cell Reports reveals1. Researchers in the United Kingdom identified an enzyme, which they called Pap1, in a strain of Pseudomonas aeruginosa isolated from a wound. They found that the enzyme can break down a plastic that is commonly used in health care because of its biodegradable properties, called polycaprolactone (PCL). Until now, the only enzymes shown to break down plastics were found in environmental bacteria, says study co-author Ronan McCarthy, who researches bacteria and bacterial infections at Brunel University in London. He says that finding the same ability in pathogens often present in hospitals could explain why these microbes persist in these environments. “If a pathogen can degrade plastic, then it could compromise plastic-containing medical devices such as sutures, implants, stents or wound dressings, which would obviously negatively impact patient prognosis,” he adds. Enzyme expression As part of their study, McCarthy and his colleagues inserted the gene that codes for the enzyme into Escherichia coli bacteria. When they made the bacteria express the enzyme, they found that it broke down the PCL in agar or beads. The strain of P. aeruginosa taken from the wound was also able to digest PCL agar and beads. And when the researchers created a mutant where the gene that coded for the Pap1 enzyme was deleted from the P. aeruginosa, the bacterium was unable to degrade the plastic. The team also found that the enzyme increased the amount of biofilm formed when P. aeruginosa was exposed to the plastic as compared with glass beads. When bacteria produce high levels of biofilm, antibiotic resistance can increase, leading to difficult-to-treat infections. The team also conducted experiments in greater wax moth (Galleria mellonella) larvae and found that the bacterium was more harmful when a PCL implant was present, compared to when the implant was absent. Additionally, moths infected with bacteria lacking the enzyme had similar survival rates regardless of whether a PCL implant was present or not. The findings highlight the increasing threat to human health of P. aeruginosa and five other highly virulent bacteria, sometimes called ESKAPEE pathogens, says Steven Djordjevic, who researches pathogenic bacteria at the University of Technology Sydney in Australia. “ESKAPEE pathogens are notorious for persisting in health-care settings and hospitals”, and for being resistant to clinically important antibiotics, he says. The authors’ analysis suggests that other ESKAPEE pathogens can degrade plastics, too, which is concerning, he adds. McCarthy says the team is developing tests to screen pathogens for the ability to break down plastics.

Symplocarpus renifolius emits sulfur-containing compounds that make it smell of rotting meat.Credit: Kirsanov Valeriy Vladimirovich/Shutterstock With a nickname like ‘skunk cabbage’, the infamously stinky flowers of Symplocarpus renifolius are unlikely to grace a bridal bouquet. But for the beetles and flies that pollinate the plant, the noxious perfume of rotting meat is an irresistible draw. Now, a sweeping study of smelly plants has revealed how skunk cabbage and several other plants concoct their malodorous aromas. The results, published today in Science1, also show that blooms that reek of faeces, rotting meat and other things that humans find off-putting are surprisingly popular in the plant kingdom. Species in one plant genus evolved the ability to make such smells repeatedly in less than 7 million years — the blink of an eye in evolutionary terms. The study also highlights how much more there is to learn about smelly plants, says Robert Raguso, a chemical ecologist at Cornell University in Ithaca, New York, who was not involved in the study. “One recent realization has been that stinky flowers are more complicated than we thought they were,” he says. “There is still a lot of discovery to be made.” Follow your nose The project began when Yudai Okuyama, a biologist at the National Museum of Nature and Science in Tsukuba, Japan, launched a survey of one of the most diverse plant groups in Japan: the genus Asarum. He and his team wanted to learn more about how the drive to court different pollinators might have influenced this diversity. Unpicking the link between smell and memories During his survey, Okuyama noticed that some Asarum species emit the odour of rotting meat. He and his colleagues decided to harness Asarum’s diversity to learn more about how plants produce these stinky compounds. They catalogued the volatile chemicals made by the flowers of 53 Asarum species and then looked for differences in gene and enzyme activity between stinky and non-stinky Asarum species. The approach pinpointed a class of enzymes called disulfide synthases, which convert certain sulfur-containing molecules into compounds responsible for the flowers’ stench. Okuyama and his colleagues found that similar enzymes had evolved independently in two other plant genera, Symplocarpus and Eurya. They also found that a change of merely three amino acids could convert an enzyme common in plants and animals into disulfide synthase. Other flora use different pathways to produce foul-smelling flowers. For example, the legendary ‘corpse flowers’ Rafflesia arnoldii and Amorphophallus titanum, which produce two of the largest flowers in the world, both smell like rotting flesh — but neither contains disulfide synthases. Stinky variety Okuyama hopes to tackle how these and other pungent flowers produce their trademark stench. These results, Raguso says, will feed into ongoing studies of how varied noxious floral odours can be. A flower might emit different compounds at different times — for example, using one composition to lure a pollinator in search of a meal or a mate, and then shifting its perfume to drive the pollinator away so that it will deposit pollen on other flowers. And flowers have been found that mimic the odour of a range of pollinator diets, such as insect ‘blood’ and rancid cheese. It’s a far cry from the long-standing, simpler idea that malodorous blossoms merely mimic faeces, or rotting flesh, fungi or fruit, says Raguso. “That gave us four ways to smell bad,” he says. “But in fact, there are many more.”

The US National Institutes of Health (NIH) plans to cut off funding for certain projects about climate change, greenhouse gases and fossil fuels. Specifically, the agency is poised to no longer financially support research on why Earth’s climate is changing, on how to enhance climate-change literacy, on assessing climate-change anxiety and more, according to internal NIH documents and e-mails that Nature has obtained. Trump gutted two landmark environmental reports — can researchers save them? The documents, which were sent to the top grants officials at 9 of the NIH’s 27 institutes and centres, provide an inside look at how the agency under US President Donald Trump will fund research at the inter of human health and climate. For those who feared that the administration might cancel all projects mentioning climate change — which Trump has in the past called a “hoax” — the latest directive comes as welcome news. But others see the NIH’s changing focus as undermining its stated commitment to extending the lifespan of people in the United States. If the goal “is to have Americans as uninformed as possible about the challenges we’re facing, that’s a pretty devastating comment on the administration’s concern for the health of Americans”, says Kristie Ebi, a specialist in climate, weather and human health at the University of Washington in Seattle. The NIH, which is the world’s largest public funder of biomedical research, did not respond to Nature’s queries about the new guidance, the forthcoming cuts to climate-related grants or scientists’ frustrations. It is unclear whether the documents have been finalized or whether they are subject to further change. Decisions, decisions Before 2022, the NIH had been investing about US$10 million per year in research on the human-health impacts of a rapidly warming world — small change for an agency with an annual budget of about $47 billion. But that year, the administration of former US president Joe Biden, Trump’s predecessor, launched the Climate Change and Health Initiative (CCHI), which prioritized environmental-health research and coordinated it across the entire agency. The CCHI received $40 million from the US Congress in both 2023 and 2024. The initiative’s fate has been in question since Trump took office in January, given that its website has been taken down, and media reports emerged that the agency would end all future research on the health effects of climate change. It now appears that the CCHI is getting rebranded: it will be called, at least internally, the Health and Extreme Weather initiative. A decision tree circulating amongst NIH staff members that indicates which climate-related grants can still be funded (see Supplemental Info below for PDF). The research the initiative funds will change too, reflecting a push to focus on extreme-weather events rather than the changing climate or the conditions making those events more likely. Nature obtained a decision tree (see above) that will be used to help grants-management staff to evaluate whether to fund or terminate projects. Much of the research that the CCHI funded, focused on the health effects of weather-related events such as wildfires or heatwaves, will continue. That comes as a relief to Ebi: “In the short term, it doesn’t matter if a wildfire is attributed to climate change — people need answers to its impacts on health and how to mitigate that risk,” she says. But a slew of research is no longer deemed within NIH’s remit, the documents show. For example, projects creating educational modules for classrooms on climate change are out. And so is one investigating how to make a more climate-friendly asthma inhaler (inhalers usually release a greenhouse gas when used). “Greenhouse gases are the reason we have climate change, and the problem will get worse unless we address them,” says Preeti Jaggi, an infectious-diseases physician at Emory University in Atlanta, Georgia, who also studies the sustainability of the healthcare industry. Joshua Rosenthal, an environmental-health researcher who co-led the CCHI until his retirement in March, says the more that Trump administration officials “narrow the ability of NIH to support research that will protect people from the harms of climate change, the more people will suffer”. Increasing anxiety Perhaps most troubling, Ebi says, is that the agency plans to halt all projects evaluating climate change anxiety, which is rapidly growing among younger populations. A 2021 survey found that more than half of young adults felt sad, anxious, powerless or had other negative emotions about climate change. “It exists, it’s happening, and you can’t erase it by not studying it,” Jaggi says. Exclusive: NIH to end billions of dollars in foreign research grants Susan Clayton, a social psychologist at the College of Wooster in Ohio who studies climate anxiety, says that these decisions will serve to increase anxiety, because avoidance and denial are linked to worse mental health. “The right way to reduce anxiety is to give people more information and to help people cope.” Under Trump, the NIH has already terminated about 800 active grants studying topics including HIV, the health of transgender people, vaccine hesitancy and COVID-19. Twelve of those grants reference climate change, according to publicly available information about the projects. Last week, the NIH also implemented a policy that will end billions of dollars in international collaborations, a move that could further hurt the agency’s ability to fund research on the health impacts of climate change, because global warming is just that — global, Rosenthal says. “It’s disappointing and harmful to erode or terminate our ability to do research on an undeniably growing set of risks,” he adds.