Collaborations with a higher proportion of unpublished researchers can yield papers that present groundbreaking ideas.Credit: Stígur Már Karlsson/Heimsmyndir via Getty What characteristics drive scientific discovery? Experience is often seen as key, with success coming from years of knowledge building and collaboration. This is certainly true for most Nobel Prize laureates, who have an average age of 58. However, research teams with high fractions of beginners — authors with no prior publication history — tend to be more disruptive and innovative, found a study posted on the preprint server arXiv on 12 September1. Actively integrating beginners into research groups could boost the disruptive and innovative qualities of the work by having fewer collaborators who are burdened by knowledge, the authors say. “Beginner scientists have less loyalty to prevailing assumptions, and they can take more intellectual freedom,” says co-author Raiyan Abdul Baten, a computational social scientist at the University of South Florida in Tampa. Can creativity in science be learnt? These researchers think so Beginner’s charm This pattern was first noticed by researchers during a study of whether artificial intelligence can predict how innovative and disruptive a paper will be in its scientific domain2. In the new study, the team analysed more than 28 million articles from SciSciNet-v2, a data lake for the science of science research. These scientific papers spanned 146 fields and were published between 1971 and 2021. Disruption scores were calculated by comparing how many citations a paper has garnered compared with the papers that it referenced. A disruptive paper that was cited more times than were the papers it referenced could, for example, disprove an established theory and offer an alternative hypothesis. During the analysis, Baten and his colleagues noticed a link between a paper’s disruptiveness and the number of beginner authors, which led them to partition data by authors’ ‘career age’ — defined as the number of years since their first publication. Beginner authors were those with no prior publications. They were surprised to find that papers with a higher proportion of beginner authors tended to be more disruptive than were papers with more early-career or senior authors. Disruption scores were highest when beginners made up the entire team, but they were also high when beginners were paired with co-authors with track records of producing disruptive work. Papers with the highest ratio of beginner authors consistently showed a positive correlation between disruption and citation count. “As the fraction of beginners increases in teams, the disruptivity and innovation go up,” says Baten. “Across team sizes, decades and disciplines, we found these results to be robust,” he adds. What makes a good PhD student? Diversity boosts scientific progress Why do collaborations with more beginners tend to be more disruptive? The researchers speculate that these scientists offer new perspectives because they are less attached to previous theories and knowledge than are their more experienced colleagues. “Sometimes, it is difficult to unlearn the prevailing assumptions then adopt radically new ones,” says Baten. This could allow early-career scientists to be more receptive to new ideas and more likely to take risks by exploring experimental approaches, he says. According to the analysis, beginners also reference more atypical papers, suggesting that, compared with more-experienced scientists, they gravitate towards work that is less commonly cited. “It certainly does fit with the idea that [newer] people might ask lots of exciting questions because they’re coming in without the baggage of knowing the framework that everybody else is working in. It could well be true that this is really important,” says Daniel Davis, an immunologist at Imperial College London. “[However], you would certainly need experience that’s able to translate a wacky idea into a set of defined experiments that are going to rigorously lead to some conclusion about that idea,” he adds. Hunter Schone, an early-career neuroscientist at the University of Pittsburgh in Pennsylvania, agrees that trainees bring fresh perspectives to projects. In August, Schone published a paper that disrupted the neuroscience field by challenging previous theories of what happens in the brain following amputation3. He doesn’t attribute the disruptive nature of the paper to his career stage, but to the hard work carried out by a team of early- and late-career scientists. However, he says that adding more senior authors to projects can make it difficult for early-career authors to be disruptive. “You have to make a lot of compromises about how work is done.” The authors hope that their study will encourage senior researchers to integrate more beginners into research groups, and therefore produce more disruptive and innovative research. “Beginners are not just blank slates or passive contributors. There’s a very interesting and systematic association with disruptive and innovative work coming from beginner heavy teams,” says Baten. Davis emphasizes the importance of considering diversity on a wider scale. “I think you do need a lot of different kinds of people to make science succeed,” he says. “Diversity of people adds to the progress of science at all levels.”

There are limited options for pain relief during pregnancy.Credit: The Good Brigade/Getty Researchers are concerned about what will happen should pregnant women follow US President Donald Trump's advice to avoid the painkiller Tylenol — also called paracetamol and acetaminophen. In an address at the White House on 22 September, Trump claimed the medication is linked to autism and other long-term neurodevelopmental conditions in the developing child. But experts warn that avoiding the drug when it is needed could place women and their fetus at even greater risk. Paracetamol is one of the most widely used drugs during pregnancy and is generally considered safe. It is commonly used to relieve pain and reduce fevers, including those caused by viral infections such as the common cold and bacterial infections that cause urinary-tract or kidney conditions. On Monday, the US Food and Drug Administration (FDA) announced that acetaminophen products would have a new warning label that lists a possible association between taking the drug and an increased chance of attention deficit hyperactivity disorder (ADHD) and autism in children. FDA commissioner Marty Makary said in a statement that the change “may lead many to avoid using acetaminophen during pregnancy”. Trump was more direct: “Fight like hell not to take it,” he said. The advice has been rejected by physicians, medical organizations and health regulators in many countries, including the American College of Obstetricians and Gynecologists, the UK Medicine and Healthcare products Regulatory Agency and the Australian Therapeutic Goods Administration. The World Health Organization said in a statement that there is no “conclusive scientific evidence confirming a possible link”. The drivers of ADHD and autism are complex and genetics plays a large part. Untreated fever But there is evidence that leaving a fever untreated, particularly if it is high (above 39.1 °C) or lasts longer than 24 hours, raises the risk of negative outcomes during pregnancy, says Debra Kennedy, a women’s health researcher at the University of New South Wales (UNSW) in Sydney, Australia. “It’s been associated with an increased risk of miscarriage and some birth defects,” she adds, including spinal and heart defects and abnormal facial features. Women should not have to be in discomfort from pain or a fever unnecessarily when there is no convincing evidence that taking paracetamol will cause harm to their child. There is some evidence that untreated high temperatures during pregnancy might increase the chance of autism and psychiatric disorders such as schizophrenia, because they can interfere with fetal brain development1. Kennedy warns that raising fear around paracetamol use could also lead women to use pain relief that is not as safe. "I definitely think there will be some women who will be reluctant to take paracetamol during pregnancy because of these statements," she says, adding that other doctors have begun asking her for resources to provide to women expressing concerns about taking paracetamol during pregnancy. Limited options Other pain-relief options have higher and well-established risks of harm when taken during pregnancy, says Alex Polyakov, an obstetrician and researcher at the University of Melbourne, Australia. These include aspirin, ibuprofen, naproxen, diclofenac and indomethacin, a class of painkillers called non-steroidal anti-inflammatory (NSAID) drugs. Studies have linked the use of NSAIDs in the first trimester to an increased risk of miscarriage2. Taking those drugs during the third trimester can affect fetal kidney function and reduce the amount of amniotic fluid around the baby, increasing the risk of poor lung development or stiff joints. Using NSAIDs during later pregnancy can also cause the early closure of an opening between two blood vessels from the heart which can lead to life-threatening breathing issues after the baby is born3. Some of these alternative painkillers are still prescribed during pregnancy, but often at lower doses than usual, says Kennedy. For example, low-dose aspirin is prescribed to reduce the risk of pre-eclampsia, which causes high blood pressure and can reduce the amount of oxygen and nutrients the fetus receives. One big issue is a lack of drug trials that include pregnant people, meaning that it is unknown whether many medications are safe to take during pregnancy. This makes some clinicians hesitant to prescribe some drugs and that women sometimes stop or avoid medications they need or terminate their pregnancy out of fear they are harming the fetus. “Including pregnant women in well-designed clinical trials would help us better understand how drugs affect both mother and child, and would address long-standing gaps in evidence,” says Polyakov.

Credit: AFP via Getty China has revealed its goal for slashing greenhouse gas (GHG) emissions, providing a glimpse into how global emissions might change over the next decade. In a video address to the United Nations Climate Summit on 24 September, Chinese president Xi Jinping announced that China will reduce greenhouse-gas emissions by 7% to 10% from peak levels by 2035. The pace at which China cuts emissions will have profound global impact. The country has accounted for 90% of the growth in the world’s CO₂ emissions since 2015 and it is now the largest GHG emitter in the world, responsible for around one-third of the global total, according to the Asia Society Policy Institute, a think tank based in New York City. Analysts have warned that China’s action could make or break the 2015 Paris agreement. In 2020, Xi pledged that China’s CO₂ emissions would peak before 2030 and that the country would achieve carbon neutrality before 2060. Some researchers say China’s CO₂ emissions will probably peak soon if they haven’t already. The latest targets are part of China’s new Nationally Determined Contribution (NDC), a climate-action plan that all countries subject to the Paris agreement must submit to the UN every five years. China also set clean-energy targets for 2035. The importance of China’s latest NDC is that its targets cover the years until 2035, past the country’s proposed peak, says Yao Zhe, a Beijing-based researcher of China’s climate policy at Greenpeace East Asia. “This is the first time China has officially outlined its post-peaking plan,” Yao says. Once China’s emissions drop, global emissions will likely start to drop, says Belinda Schäpe, a China analyst at the Centre for Research on Energy and Clean Air (CREA), a Helsinki-based think tank. “That's why these targets are so important for the global community, because they can help them understand” how the world’s emissions trajectory could look, she says. This is also the first time that China has announced a target that covers not only carbon dioxide (CO₂) but all GHGs including methane and nitrous oxide, says Zhang Da, who researches energy economics and climate change at Tsinghua University in Beijing. Ambitious or not? Some researchers think that China’s emissions-reduction target falls short of what the world needs to achieve the Paris agreement’s aim, of limiting global warming to well below 2 °C above pre-industrial levels, and striving to stay below 1.5 °C. “Anything less than 20% is definitely not aligned with 2 degrees. Similarly, anything less than 30% is definitely not aligned with 1.5 degrees,” says Lauri Myllyvirta, an analyst who has tracked China’s emissions trends for more than a decade and is CREA’s co-founder. Myllyvirta cites his analysis of a set of future climate scenarios used by the Intergovernmental Panel on Climate Change to help the world adhere to the Paris Agreement. The way China has defined its emissions cuts — as 7–10% of an undefined amount, rather than specifying a year as the basis for calculation – leaves the door open for short-term emissions increases, Myllyvirta says. The different pathways for China to achieve carbon neutrality between 2030 and 2060 could result in different amounts of cumulative emissions, says Myllyvirta. “What matters for the climate is the total amount of GHGs emitted into the atmosphere over time,” he says, adding that this is why cutting emissions fast early on is important. But others, such as Da, regard China’s target an important step. “Reducing non-CO₂ emissions is typically more challenging than mitigating CO₂,” says Da. “A 7–10% reduction in net GHG emissions from peak levels usually implies a higher level of CO₂ reduction.” A study by Da and colleagues1, published in January, found that if China reduces its energy-related CO₂ emissions by 10–12% from peak levels by 2035, the country would meet its goal of being carbon neutral before 2060. A separate study2, also co-authored by Da, found the two-degree temperature goal is achievable under China's carbon-neutral timeline. Different logic To some, China’s emissions target did not come as a surprise. “Chinese policymakers often underpromise on climate targets so they can overdeliver, prioritizing the implementation of commitments over ambitious numbers on paper,” says Norah Zhang, an analyst at the NewClimate Institute, a non-profit organization in Berlin. For example, China achieved its 2030 NDC targets of installing 1,200 gigawatts of wind and solar capacity six years ahead of schedule. The rationale behind the use of targets is deeply tied to the country’s top-down political system, according to Yao. “Setting and evaluating targets is a key means through which the central government manages the country. As a result, there is a strong political culture of taking targets seriously,” Yao says. Policymakers usually take a realistic approach to setting targets, she adds. Although China’s target seems modest in percentage terms, in absolute terms “this is huge”, says Piers Forster, a climate physicist at the University of Leeds, UK. In his estimation, the GHG emissions that China will need to cut by 2035 would be equivalent to “three UKs completely decarbonizing over the next decade”.

Autism research funded by a new US initiative includes study of environmental exposures during pregnancy. Credit: Diana Bagnoli/Getty Scientists moved a step closer to understanding the complex causes of autism this week. Although all of the headlines went to US President Donald Trump’s poorly evidenced statements that the painkiller acetaminophen is linked to the neurodevelopmental condition, his White House autism event brought welcome — and largely overlooked — news to scientists: the US National Institutes of Health (NIH) is investing US$50 million in an unusual autism-research effort. Trump and Jayanta Bhattacharya, director of the NIH, announced on 22 September that 13 research groups will receive funding under the Autism Data Science Initiative (ADSI), a Trump administration programme to fund studies that explore how interacting genetic and environmental factors contribute to autism. “This is where the field needs to be going in searching for the complex causes of autism,” says Helen Tager-Flusberg, who studies autism at Boston University, Massachusetts. Trump team backs an unproven drug for autism — but does it work? The funded projects range from studies on environmental exposures during pregnancy to experiments on brain cells. Funding was also awarded to efforts to replicate the projects’ results and so ensure that they are robust. Researchers, although pleased by the aims of the funded work and the rigour of the methods, have some concerns about the project. Several ADSI-funding recipients say that they are expected to complete their projects relatively quickly — within three years instead of the usual five — and some say that they are alert to political interference with their results. Trump prompted fierce pushback from scientists with his statements about acetaminophen earlier this week, given the lack of convincing evidence to support a link with autism. “We should wait until the research happens before announcing an answer,” says Jason Stein, a neuroscientist at the University of North Carolina, Chapel Hill, who received an ADSI grant. “This is not political interference, but rather a bold, science-driven effort to deliver meaningful answers more quickly,” said a spokesperson for the US Department of Health and Human Services (HSS), which oversees the NIH. Quick turnaround The NIH announced the ADSI in May and invited researchers to submit grant applications for research into the causes of autism, its growing prevalence and potential interventions. Some researchers expressed concern that applicants had only a month to submit proposals — much less time than usual — and it was unclear who was reviewing the grants and with what criteria. Some worried that the funding would be channelled to researching the discredited idea promoted by Trump’s health secretary Robert F. Kennedy Jr that vaccines are linked to autism. “Some people thought: maybe we should steer clear of this,” says Judith Miller, a psychologist who studies autism at the Children’s Hospital of Philadelphia in Pennsylvania. In the end, nearly 250 research teams applied, and no awards were granted to projects that focus explictly on autism and vaccines. Several of the projects will involve exposomics: the study of the array of environmental factors to which a person is exposed. Miller is leading a three-year, $4.3-million project combining genome and exposome data to seek factors associated with autism. The project will draw on previously collected data on more than 100,000 children, including about 4,000 autistic children, and connect those to maternal-health records. The research team plans to use information on where participants live to add in data on air quality, access to green spaces and other environmental markers. “We haven’t been able to bring this type of data all together in a clinical population,” before, Miller says. Replication requirement Stein and his team, by contrast, are examining autism using brain organoids grown from the stem cells of autistic and non-autistic children. The researchers plan to expose the tissue to substances that epidemiological studies have linked to autism — such as valproic acid, a drug used to treat epilepsy — and examine how this affects gene activity. Trump links autism and Tylenol: is there any truth to it? The team expects to be asked by the NIH to look at acetaminophen or other substances too, says Joseph Piven, a psychiatrist at University of North Carolina, Chapel Hill, who is also working on the organoid project. “As long as they have some detectable level of epidemiological evidence, I think that’s a valid question to go forward,” he says. The ADSI is building in replication efforts from the start. Judy Zhong, a population-health scientist at Weill Cornell Medicine in New York City, has received around $5 million from the ADSI for a centre that will require other ADSI-funded investigators to hand over their computer models so that their results can be independently replicated. “It is very unusual,” Zhong says. Collaborative approach But researchers are still worried about political interference in autism research. Some point to the announcement earlier this month that the HHS would award a contract to Rensselaer Polytechnic Institute in Troy, New York, to search for an association between vaccines and autism in databases. “Is this the best use of funds to support another investigation, on what appears to be a largely settled question?” says Craig Newschaffer, an autism researcher at Pennsylvania State University in University Park. Some researchers would like to see more funding for research that helps autistic people to lead healthy and fulfilling lives — a primary focus of only 2 of the 13 ADSI grants. Katharine Zuckerman, a paediatrician at Oregon Health & Science University in Portland, will be using her $4.25-million grant to look for factors in childrens’ lives — such as regular doctor’s visits or attending quality schools — that correlate with outcomes that autistic people say are important to them, such as sleep or good mental health. Like the other ADSI projects, this will be done in consultation with the autism community. “Looking at the cause of autism is important, but I think that it’s also important that we address the concerns of autistic people who are here today and what we could do to improve their lives,” Zuckerman says.

Mitochondria (illustration) have their own DNA, which they expel into their environment if the molecules don’t meet their standards.Credit: Kateryna Kon/Science Photo Library The cellular batteries known as mitochondria sometimes dump DNA into their surroundings, which can contribute to inflammation during ageing. Now a study in mice reveals why this dumping occurs: mitochondria are expelling ‘tainted’ DNA1. Scientists found that, in the cells of ageing mice with kidney inflammation, strands of mitochondrial DNA (mtDNA) contained an excess of certain types of nucleotides — molecular building blocks — that can harm DNA. This excess prompted the mitochondria to eject the abnormal fragments of genetic code into the cytosol, a fluid that fills the cell, in which the free-roaming mtDNA kick-started key inflammatory pathways associated with ageing. The study is exciting because it helps to explain why and how mitochondria throw away their DNA, says Timothy Shutt, a medical geneticist at the University of Calgary in Canada, who focuses on mitochondria. This insight could help researchers to better understand mitochondria’s contribution to inflammageing — the chronic inflammation that occurs as people get older, adds Shutt. The findings were published 24 September in Nature1. Out with the rubbish Mitochondria are energy-producing organelles that have their own genome. When the mtDNA is damaged or disrupted, the mitochondria kick it out, sending it into the cytosol. This can occur when the relative levels of certain nucleotides in the mtDNA become too high or low. This oversupply problem, which stresses the mitochondria, can be caused by certain drugs and also seems to happen in ageing cells. But exactly how this nucleotide imbalance leads to mtDNA release and inflammageing had been unclear, says study co-author Thomas Langer, a cell biologist at the Max Planck Institute for Biology of Ageing in Cologne, Germany. How quickly are you ageing? What molecular ‘clocks’ can tell you about your health To find out, Langer and his colleagues turned to mice engineered to lack an enzyme called MGME1. As these mice age, their kidneys usually become inflamed, making them useful models for studying inflammation. MGME1 ensures that the mitochondrial genome makes accurate copies of itself, but the connection between its loss and inflammation was unclear. The researchers detected loose mtDNA fragments in the kidney cells of the enzyme-lacking mice, but not in controls. The loose pieces of genetic material bound to and turned on an enzyme that’s a known contributor to inflammation in aged tissues. This confirmed that free-floating mtDNA is a key driver of inflammation when MGME1 is missing. Fussy, fussy But what triggered the mitochondria to throw away the mtDNA in the first place was unclear. When the researchers took a closer look at the kidney cells of the modified mice, they found that the cells contained relatively low levels of DNA building blocks called deoxyribonucleotides. That short supply forced the mitochondrial DNA to incorporate unusually large numbers of RNA building blocks while it made copies of itself. This excess of the “wrong” kind of building block hinders DNA replication, a process that the lack of MGME1 probably exacerbates further, says Langer. This could explain why the organelles expelled the mtDNA into the cytosol, triggering inflammation. The findings help to answer a long-standing question about why mtDNA leaks out of mitochondria and drives inflammation, says David Sinclair, a geneticist at Harvard University in Cambridge, Massachusetts. But more work needs to be done to determine whether this process occurs naturally during ageing or whether it happens only under specific conditions. “The big questions is — is this process relevant to normal physiology?” says Sinclair. The next step for Langer and his colleagues is to better understand how discarded mtDNA contributes to cellular ageing and inflammation. “This is certainly an open question,” says Langer.

Download the September 2025 long read podcast In April, Robert F. Kennedy Jr held a press conference about rising diagnoses of autism, and said he would soon be announcing a study to find the responsible agent. Although Kennedy said that environmental factors are the main cause of autism, research has shown that genetics plays a bigger part. Also, the rise in prevalence, many researchers say, is predominantly caused by an increase in diagnoses rather than a true rise in the underlying symptoms and traits. Although the US National Institutes of Health (NIH) announced a US$50 million to fund studies on the causes of autism, many researchers were dismayed that these developments seemed to ignore decades of work on the well-documented rise in diagnoses and on causes of the developmental condition. This is an audio version of our Feature: Autism is on the rise — what’s really behind the increase? Never miss an episode. Subscribe to the Nature Podcast on Apple Podcasts, Spotify, YouTube Music or your favourite podcast app. An RSS feed for the Nature Podcast is available too.

A brain scan of a person with Huntington’s disease, which causes a loss of brain volume as neurons are killed off by the accumulation of a mutant protein.Credit: Zephyr/Science Photo Library A one-time gene therapy can markedly slow the progression of Huntington’s disease, potentially paving the way for the first ever treatment to alter the course of this rare, inherited brain disorder. In a small trial of 29 people who were in the early stages of Huntington’s-related decline, participants who received a high dose of the therapy directly into their brains saw the disease slow by 75% over three years, compared with those in a control group. Genetic therapies offer new hope against incurable brain diseases The benefit was statistically significant across several clinical measures, according to data released this week by uniQure, a gene-therapy company based in Amsterdam. Trial investigators also observed a reduction in the level of a toxic protein linked to neurodegeneration in the spinal fluid of people who received the therapy. On the strength of these findings, uniQure executives said they plan to seek regulatory approval for the treatment next year. “This gene therapy is obviously a big step forward,” says Sandra Kostyk, a neurologist at the Ohio State University Wexner Medical Center in Columbus, who was involved in the trial. “The data look quite good.” Slowing the disease’s advance could translate into many extra years of independence for people with Huntington’s, Kostyk says, but it is not a cure. And, with so few participants, the trial’s results — still unpublished — should be viewed as preliminary, she adds. “I think we need more time and more data.” A deadly repeat People living with Huntington’s disease typically see their symptoms progress year by year, usually starting between the ages of 35 and 55. What often begins as a subtle loss of coordination or forgetfulness then often progresses to involuntary movements, sharp mood swings and a gradual unravelling of memory and thought. The disease is caused by excessive DNA repeats in a gene called huntingtin, which leads to the production of a faulty protein that slowly poisons the brain. There are currently no therapies that address this root cause, so those who inherit the mutation are left only with drugs that ease symptoms. Some of the first attempts to develop a treatment focused on antisense therapy, a gene-targeted strategy that uses short strands of DNA or RNA to dial down production of the defective huntingtin protein. The approach showed promise in early clinical development1. But hopes dimmed in 2021 after a leading candidate failed in late-stage testing, with those receiving the therapy seeming to fare worse than those who were given a placebo2. This clinical setback redirected attention towards a different strategy: gene therapy, which aims to provide a one-time intervention that permanently silences or modifies the deficient gene at its source. Molecular muzzle In the case of uniQure’s gene therapy, the treatment uses a harmless virus to deliver the recipe for making a short RNA sequence known as a microRNA directly into cells in the affected parts of the brain. The microRNA is designed to ‘muzzle’ the defective huntingtin gene — and stop the cells producing the faulty protein — by blocking the molecular instructions encoded by the gene, known as mRNA. Once delivered, the virus-encoded instructions stay inside the cells, which continue to produce the therapeutic microRNA. The discovery of microRNAs was feted with a Nobel Prize last year, although the technology has yet to yield any approved medicines. Administering the treatment requires a lengthy surgery in which clinicians use magnetic resonance imaging to precisely place a cannula through small holes in the skull. The therapy is then infused slowly into the striatum, a part of the brain that is among the first and hardest hit by Huntington’s disease. What’s so special about the human brain? A graphical guide “It’s not a minor procedure,” says Kostyk. And although most people tolerated the therapy itself — with none of the main safety red flags that have plagued gene therapies for other brain conditions — some individuals experienced headaches, pain and other complications from the surgery. A rival gene-therapy programme with a similar molecular design and delivery strategy — developed by biotechnology company Spark Therapeutics, which is now owned by pharmaceutical giant Roche in Basel, Switzerland — also entered clinical testing earlier this year. Both treatments, if approved, are expected to cost more than US$1 million per person, a price tag that would put them out of reach for many people and strain health-care systems worldwide. Similar issues of access and affordability could also shadow what many researchers see as the next leap forwards in the treatment for Huntington’s disease: interventions built around CRISPR and other gene-editing technologies that could offer a permanent cure. But, practical constraints aside, the apparent ability of uniQure’s therapy to slow disease progression by three-quarters is prompting Huntington’s researchers to imagine a future in which a fatal genetic sentence becomes a treatable condition. “This feels like it could be a turning point,” says Andrew Duker, a neurologist at the University of Cincinnati in Ohio, who is involved in the Spark Therapeutics trial. The results are “the first step in really demonstrating that Huntington’s disease can be slowed down”, adds Kyle Fink, a neuroscientist at the University of California, Davis. “It’s going to really set up Huntington’s to be a good candidate for next-generation treatments.”

A crime laboratory officer removes material from the house of the prime suspect in the Gilgo Beach killings.Credit: Jeenah Moon/AP Photo/Alamy A judge in New York rejected a request on 23 September to disqualify the use of cutting-edge DNA sequencing as evidence in a case against an alleged serial killer. The ruling paves the way for a type of DNA analysis known as whole-genome sequencing — used to decipher ancient DNA in fossilized remains, for example — to be admitted as evidence in US criminal trials. “This is a big step forward for the use of DNA evidence, because it will allow comparisons and matches with evidence that was previously considered too minuscule, too old or too badly degraded to be considered useful,” says Nathan Lents, a biologist at the John Jay College of Criminal Justice in New York City. “You can bet that cold-case units all over the country are going back through their evidence lockers to see if there are samples that can now be tested with a reasonable chance of success.” A technology matures The Gilgo Beach killings were a series of murders that occurred in Suffolk County, New York, between 1993 and 2011. Many of the victims were sex workers, and the bodies of some of them were found near the eponymous beach. New York architect Rex Heuermann was charged with seven of the murders in 2023 and 2024 partially on the basis of hairs found on the victims. In 2010, investigators concluded that there wasn’t enough genetic material in the hairs for conventional forensic technology to identify whose they were. But since then, genomic technology has matured at a fast pace. Rex Heuermann appears in a New York court in 2023. Credit: James Carbone/AP Photo/Alamy Forensic scientists still identify suspects from hair and other samples using the conventional polymerase chain reaction (PCR) method, which can amplify snippets of DNA called short tandem repeats (STRs). Today, however, scientists can also sequence whole genomes and look at single-letter variations in the DNA to zero in on a person of interest. This technology has the added benefit of being able to analyse severely degraded samples — from Neanderthal remains to hairs in the Gilgo Beach case. But it has typically been used only during investigations, and the results have not been widely presented as evidence before juries. The controversial company using DNA to sketch the faces of criminals The prosecution in the Gilgo Beach case contracted the company Astrea Forensics in Santa Cruz, California, to analyse the hairs found on the victims. Because they were rootless, they hadn’t yielded enough DNA for conventional STR profiling. (Rootless hairs are mostly made of keratin and don’t contain much genetic material, Lents says.) But Astrea was able to sequence the small amount of genetic material in the hairs and then use statistical-analysis software to scan the DNA for single-nucleotide polymorphisms (SNPs) — single-letter differences in people’s genetic codes. Even in a degraded sample, whole-genome sequencing followed by this statistical analysis can spot thousands of SNPs: enough to pinpoint an individual’s unique pattern, Lents says. Next, scientists at Astrea compared the SNP profiles they obtained with profiles stored in genealogy databases and looked for matches. People submit their DNA to genealogy services, such as 23andMe and Ancestry, and these companies return SNP profiles that people can then upload to websites such as GEDMatch to locate potential relatives. Users have the option to make the data available to forensic genealogists — and these data are what Astrea used to zero in on Heuermann and his family. Setting precedent The defence team in the Gilgo Beach case initially challenged the validity of Astrea’s DNA-analysis methods — which were developed by palaeogenomics researcher Richard ‘Ed’ Green at the University of California, Santa Cruz. (Green declined to comment for this article, given the ongoing prosecution.) The defence argued that the statistical methods used by Astrea were unique, and questioned whether they had been sufficiently validated. What went wrong at 23andMe? Why the genetic-data giant risks collapse But on 3 September, the judge ruled during a special hearing that the prosecution could bring this evidence before a jury. This is the first time that the technology has cleared a ‘Frye hearing’, during which “you have to show that a technique is accepted by the scientific community”, says David Gurney, director of the Investigative Genetic Genealogy Center at Ramapo College of New Jersey in Mahwah. This kind of hearing is used as an evidential standard in other US states, which is why the Gilgo Beach case could establish a legal precedent in the United States, with other countries potentially following suit. Gurney says similar techniques have been used during US criminal investigations to identify suspects, including the Golden State Killer, but conventional DNA forensics methods were also used. (The Golden State Killer was responsible for a string of burglaries, assaults and murders in California in the 1970s and 1980s.) When such cases have reached the courts, it was the STR profile that was presented. There have been a handful of cases in which SNP analysis was presented to the jury, including a trial in Idaho last year, but the judge did not require the stringent evidential hearing that happened this month in New York — so those cases did not set a precedent. Scrutinising the evidence Pulling the family members of criminals into an investigation, or anyone whose hair happens to end up at a crime scene, is a risk of SNP-based forensic genealogy, Gurney acknowledges. But he says the technique also has the potential to narrow the list of suspects so that many others don’t have to have their lives and privacy disrupted by the police. Legal specialists consulted by Nature for this story, however, say that STR profiling, which is cheaper and faster than whole genome sequencing, is likely to remain standard in most cases where there is sufficient physical evidence to perform it. Supercharged crime-scene DNA analysis sparks privacy concerns This week, the judge rejected the defence team’s second attempt to get Astrea’s results thrown out on the basis that the California company is not licensed by the New York health department. The prosecution prevailed in their counter argument that forensic laboratories are not required by state law to have such a licence. Lents says he expects Heuermann’s defence to continue challenging the use of advanced DNA analysis in the trial, but that’s not a bad thing. “Our adversarial system puts a great deal of pressure on the government to make its case and a great deal of scrutiny on the methods used to process and analyse evidence,” Lents adds. This standard, and the work of defence attorneys — who challenge the evidence at every step — “is about protecting all of us”, he says.

After a mouse received treatment to eliminate immune cells called microglia, it was injected with human progenitor cells that developed into human immune cells (green, pink and blue) in the animal’s brain.Credit: M. M.-D. Madler et al./Nature A fresh supply of the immune cells that keep the brain tidy might one day help to treat a host of conditions, from ultra-rare genetic disorders to more familiar scourges, such as Alzheimer’s disease. In the past few months, a spate of new studies have highlighted the potential of a technique called microglia replacement and explored ways to make it safer and more effective. “This approach is very promising,” says Pasqualina Colella, who studies gene and cell therapy at Stanford University School of Medicine in California. “But the caveat is the toxicity of the procedure.” New hope for Alzheimer’s: lithium supplement reverses memory loss in mice Microglia are immune cells that patrol the brain, gobbling up foreign invaders, damaged cells and harmful substances. They can help to protect neurons — cells that transmit and receive messages to and from other tissues — during seizures and strokes, and they prune unneeded connections between neurons during normal brain development. “Microglia do a lot of important things,” says Chris Bennett, a psychiatrist who studies microglia at the Children’s Hospital of Philadelphia in Pennsylvania. “So, it’s not surprising that they are involved in the pathogenesis of many diseases.” Those diseases include a suite of rare disorders caused by genetic mutations that directly affect microglia. Malfunctioning microglia have also been implicated in more familiar conditions with complex causes, such as Alzheimer’s disease and Parkinson’s disease, as well as ageing, says Bo Peng, a neuroscientist at Fudan University in Shanghai, China. Immune-cell swap This has led researchers to investigate a tantalizing possibility: that replacing disease-causing microglia could treat some brain conditions. But swapping out microglia poses special challenges. Physicians typically replace a person’s immune cells by performing a bone-marrow transplant, which provides a fresh supply of stem cells that shelter in the bone marrow and give rise to many immune cells. Microglia, however, reside almost exclusively in the central nervous system, and typically replenish themselves by division rather than relying on stem cells in the bone marrow to send in replacements. How CRISPR gene editing could help treat Alzheimer’s Physicians already use bone-marrow transplants to treat some rare diseases that affect microglia, such as a condition called X-linked adrenoleukodystrophy. The treatment can be effective, says Marco Prinz, a neuropathologist at the University of Freiburg in Germany, but the results are inconsistent and typically replace only a small percentage of the recipient’s natural microglia. In July, Peng’s team used bone-marrow transplants to replace abnormal microglia resulting from a fatal brain disease called CAMP (CSF1R-associated microgliopathy). The treatment was a success in both mice and in a small trial of eight people with the rare disorder: none of the eight participants experienced a decline in their motor or cognitive abilities in two years following the treatment1, whereas members of a control group who did not receive the procedure experienced a deterioration of both. One possible reason for the success of the CAMP trial is the nature of the disease itself, says Bennett, because people with CAMP tend to produce relatively few microglia. This could leave space for the transplanted cells to thrive. Onerous regimen Creating that niche for the new microglia is a pivotal step in microglia replacement — and a source of concern. To make room for transplanted cells, physicians must first wipe out as many of the brain’s resident microglia as possible. This can entail high levels of chemotherapy or radiotherapy, both of which can leave the recipient vulnerable to infection during the procedure and raise their long-term risk of cancer. That means that microglia replacement is, for the moment, too toxic to be used except in severe, rapidly progressing diseases such as CAMP, says Colella. CRISPR helps brain stem cells regain youth in mice In August, Prinz’s team reported using microglia replacement to treat mice with a version of Sandhoff disease, a condition that kills neurons2. The team chose this form of therapy after finding that a Sandhoff-causing mutation interrupts proper communication between microglia and neurons. The researchers found that using bone-marrow transplants from mice that did not carry the Sandhoff mutation improved survival and movement in mice with the disease. Also in August, stem-cell biologist Marius Wernig at Stanford University School of Medicine and his team independently used microglia replacement to treat Sandhoff disease — but without a bone-marrow transplant. First, the researchers isolated a specific population of cells that give rise to microglia and grew more of them in the laboratory. They then injected the resulting cells directly into the brains of mice3. This meant that there was no need to irradiate the entire body, so the team was able to use radiotherapy just on the heads of the mice. That approach could spare recipients some of the side effects of full-body radiation treatment, but exposing the head to radiation still raises safety concerns, says Peng. For example, such treatment could kill stem cells in the brain that give rise to new neurons, he notes. Drug couriers Other methods might yield safer techniques in the future. Earlier this year, a study found that three rounds of treatment with a microglia-killing drug might be sufficient to clear the way for transplanted cells4. Once the safety questions are addressed, microglia replacement might be used to treat complicated brain disorders with many causes, such as Alzheimer’s disease. Microglia could also be one day harnessed to carry molecules into the brain, which is surrounded by a protective barrier that keeps many drugs out. “They are like a trojan horse for the brain,” says Colella. For now, however, it makes sense to focus on genetic conditions that directly target these immune cells, says Prinz. “We start with the low-hanging fruit,” he says. “The big promise in the future, I hope, will be more complex diseases.”

If the 264 million students enrolled in higher education around the globe were a country, it would be the fifth most populous in the world. Some 53% of its citizens would identify as women and most would be located in Asia. Residents would speak and study in hundreds of languages, but English would dominate. The future of universities This nation of learners would also be one of the fastest growing. Since 2000, the number of university students around the world has more than doubled, and the number that cross borders to learn has roughly tripled, to almost seven million. Aided by the Internet, conferences, shared curricula and collaborations, the world of higher education has become more tightly linked. But the pattern of interconnected growth is beginning to unravel as wealthy Western nations become much less welcoming to foreign students. The administration of US President Donald Trump in particular has been targeting higher-education institutions and international students. Many of the latter are consequently looking elsewhere to earn their degrees, and those opportunities are growing, especially in some low- and middle-income countries. But expanding access to higher education has also raised concerns about the quality and value of that education. Universities must move with the times: how six scholars tackle AI, mental health and more These changing currents are particularly crucial for science, which relies on the infrastructure of universities to train its students, share ideas and do research. “The current global scientific system is facing unprecedented risks to the things that have made it robust,” says Chris Glass, a higher-education researcher at Boston College in Massachusetts. Specialists caution that there isn’t a universal portrait that captures the state of higher education around the world. “There’s never a global story,” Glass says. But he and other scholars have worked to identify key trends in the demographics of this ‘country’ of higher education — and predict where it’s going. Here, Nature examines some of these trends: the growth of higher education, especially in less-wealthy nations; the geopolitical forces created by and shaping higher education; and how all this affects who gets to learn and what they are taught. Who goes to university? The overarching trend in higher education over the past half century has been an explosive growth in student numbers. Higher-education researchers typically track this growth by comparing the number of enrolled students in a country with its population of student-aged people, a measure called gross enrolment ratio (GER). (This ratio can even be higher than 100% because people outside this age range can attend universities.) For countries in Western Europe and North America, participation in higher education is now the norm (see ‘Rising tides in higher education’). The higher-education GER in these regions increased from 61% in 2000 to 80% in 2024. In the past few years, however, this group has been surpassed by Central Europe, who leapt from an average GER of 42% to 87% in that same time span. The massive increase came mostly from an explosion of private education after the collapse of the Soviet Union. Source: UNESCO Other regions of the world are beginning to catch up: between 2000 and 2023, eastern and southeastern Asia’s GER went from 15% to 62% and Latin America and the Caribbean’s went from 23% to 58%. Many countries are making explicit efforts to expand their GER. India, which had a GER of 28% in 2021–22, has set a target of reaching 50% by 2035. The global average, which was 19% in 2000, has shot up to 44% in 2024. Some regions are still struggling, however. Sub-Saharan Africa had a GER of 9% as of 2021, up from 4% in 2000. It is also the only region in which women are still under-represented in higher education; there are roughly 76 women for every 100 men. The primary obstacle is funding, says James Jowi, director of the African Network for Internationalization of Education, a non-profit organization based in Eldoret, Kenya. “Most young people can’t afford a college tuition fee,” Jowi says. Universities under fire must harness more of the financial value they create Increasing access is not without concerns, however, says Laura Rumbley, the director for knowledge development and research at the European Association for International Education in Amsterdam. As peoples’ level of education rises, so too do the qualifications necessary for good jobs. In regions with extremely high enrolment, higher-education degrees, including master’s and PhDs, become more necessary and less valuable, she says. “We want to avoid higher education becoming a dead end for individuals and for societies.” Higher education should open doors, not close them, she says. Where do students go to learn? As the number of students and institutions has expanded, higher education has become more globally linked, but the shape and nature of those connections is changing. Researchers describe these connections as internationalization: how globalization takes place at university. The clearest and easiest way to see this is in student mobility, Rumbley says. In 2000, 2.1 million students crossed borders to learn; today, 6.9 million do, about 2.6% of all students. In some regions, such as the European Union, international travel for education is much more common: about 9% of its students go abroad, often to other countries in the EU. Much of the mobility comes from postgraduate students seeking more specialized education that is not necessarily available locally. And the flow of students is generally towards wealthy nations that have invested heavily in public education. For example, the United States has about 20 million students in higher education, 1.1 million of whom come from other countries. Meanwhile, India has 43 million students but only 46,000 are international (See ‘Incoming and outgoing students’). Source: UNESCO However, specialists say there might be a slowing down of this typical movement because of two factors: an increase in the availability of education in low- and middle-income countries and a decline in access to universities in many wealthy nations. The United Kingdom, Canada and Australia have all imposed immigration restrictions that have begun to chip away at undergraduate and postgraduate enrolment from abroad. And anti-immigration policies in the United States are increasing. The Trump administration has revoked more than 1,400 visas, restricted travel from 19 countries and proposed rules to limit visas for PhD candidates to 4 years, despite US programmes often taking longer than that to complete. NAFSA: Association of International Educators, a non-profit organization based in Washington DC, projects that there will be roughly 30–40% fewer international students entering the United States by 2025–26 than for the previous academic year, at a cost of about US$7 billion in revenue and 60,000 jobs. Before the start of the twenty-first century, the United States had a “diversified portfolio” of international students, Glass says, from many countries. Over the past two decades, however, China and India have come to dominate, going from about one-third to more than half of the number of students in the US international study body. This rise has been fuelled by demand from increasingly wealthy middle classes in those countries and opportunities provided by US universities hungry for more funds, Glass says. Students in China and India now have other options. Although the United States, the United Kingdom, Canada and Australia are still top destinations for international students, countries such as the Netherlands and South Korea are increasingly popular. And China is working to attract more foreign students. In 2000, the United States was host to about 26% of all international students; today that number is about 16%. How universities came to be — and why they are in trouble now Another way that universities have tried to tap into increasing global demand is through satellite campuses: extensions of highly visible institutions in foreign cities, such as New York University’s venture in Shanghai, China. These institutions have created opportunities for international education without the need to cross borders. In 2023–24, the United Kingdom hosted about 732,000 international students domestically, but it educated roughly 621,000 students abroad, studying at foreign extensions of UK universities. “That might be a trend which will be increasing, because then you have students that can say, ‘Well, I don’t have to travel abroad, but I still can have a foreign degree’,” says Hans de Wit, a higher-education researcher at Boston College. This transnational approach can avoid troublesome visa issues, but raises difficult questions, he says, such as whether it is really the same quality education? For many it is not. “If I were a student, I would be studying in Bristol University in the UK, not in Delhi,” says Saumen Chattopadhyay, an economist who studies higher education at Jawaharlal Nehru University in New Delhi. Studying abroad provides more opportunities to emigrate and gain access to international contacts. Although opportunities to go abroad can help individuals, they can drain talent and money from less-wealthy countries. “It has been considered for a long time a threat to the human capital development from African countries,” Jowi says. He is working to develop intra-Africa mobility and reduce brain drain by building local centres of excellence at universities in southern and eastern African countries. What do universities teach? Internationalization has also changed what students learn. Much like people, ideas, research and curricula can also cross borders. No lectures, exams, essays: inside a twenty-first-century university In 2010, a wide swathe of countries ranging from Iceland to Russia joined the European Higher Education Area (EHEA). The EHEA has attempted to set standards for degrees — bachelor’s, master’s and doctorates — as well as a common quality-assurance system. “There was a lot of resistance to that from the academic community,” de Wit says. But the standards have worked, and aided international collaboration between countries in the EHEA. Although precise numbers for global scholarships are hard to come by, higher-education researchers agree that universities have placed more emphasis on science, technology, engineering and mathematics (STEM) degrees than they have before (see ‘STEM PhDs rising’). China, for example, has risen to the top of many global university rankings mainly on the basis of its STEM research — which is where both economic and reputational rewards are. Source: NCSES The rise in STEM research in China has led to a cross-national partnership with the United States — the largest in the world. Researchers collaborating between the two countries produced 60,000 papers in 2020. Growth here has slowed, however, as concerns about national security and geopolitics have come to the fore. Researchers in China have also begun to depend less on US collaborators. What institutions do they attend? Much of the growth in higher education has been in the private sector. This privatization has been especially rapid in fast-growing areas such as southeast Asia and Latin America, where government funding cannot keep up with the demand (see ‘University privatization’). In Africa, roughly half of all higher-education institutions are private. Source: AISHE The rise in private-sector education has been driven by and shaped attitudes about the value of education. “Education has both public benefits and private benefits,” says Melissa Whatley, a higher-education researcher at the College of William and Mary in Williamsburg, Virginia. For countries, education is an economic boon and a key resource. For individuals, it’s an opportunity for a higher income or different social status. An increasing emphasis on the individual benefits from education has created a demand that the private sector fulfils. Researchers say that private-sector education can fill fields in which public education is not sufficient or slow to expand into, such as computer science, business administration and special areas such as divinity. However, private education tends to stay away from the most education-intensive areas, such as PhD research or medical training, Rumbley says, so it cannot completely cover shortfalls in public funding. Universities are — and must continue to be — a force for good There are also concerns about quality. Aside from a few countries, such as the United States and Japan, the quality of private higher-education institutions has typically lagged behind the public ones, in part owing to looser regulations, say researchers who spoke with Nature. Private universities in Africa have, at most, around 40% of academic staff with PhDs, Jowi says. Some countries, such as India, are trying to find a balance between public- and private-sector education. “Commercialization is inimical to the delivery of quality education,” Chattopadhyay says, but there are ways that private-sector education can efficiently match industry needs. India’s 2020 National Education Policy, for example, pairs public institutions with private partner institutions so that they can learn best practices from one another. Amid the uncertainty of a world roiled by technological and geopolitical shifts, it can be difficult to draw conclusions about the present, let alone the future. “It’s easy to day-trade in the day-to-day policy volatility” seen in the United States and other countries, Glass says. But a few things seem certain: in the coming years, higher education will keep growing. It will serve more students. Achieving equity of access and quality will still be a crucial concern. Barriers to entry will still exist, but so will opportunities. In a decade or so, as many as ten million students are projected to travel internationally for their education. “The university,” Rumbley says, “is essentially an international animal.”