A research team from The Chinese University of Hong Kong(CUHK)’s Faculty of Medicine(CU Medicine)has conducted alarge cohort study among 1,627 children with and without autism spectrum disorder(ASD)and found alterations in four kingdoms of the gut microbial species including archaea,bacteria,fungi and viruses in children with ASD.Using machine learning,they developed apanel of 31 multikingdom and functional markers that showed high diagnostic performance for ASD and has great potential as aclinical diagnostic tool.The findings were published in Nature Microbiology.In apilot study,the researchers also showed that modulation of the gut microbiome alleviated symptoms of anxiety in children with ASD,introducing the possibility of anew therapeutic paradigm for the condition.Multikingdom gut microbial markers facilitate ASD diagnosis ASD is aneurodevelopmental condition characterised by impairment in social communication,and restrictive and repetitive behaviour.Genetic and environmental factors contribute to the pathogenesis of ASD but emerging evidence suggests that impaired cross-talk between the gut microbiome and central nervous system,dubbed the gut-brain axis,may contribute to the development of ASD.According to the latest estimation from CU Medicine,approximately 2.54%of children in Hong Kong has ASD,and the incidence has been rising in recent years.The CU Medicine research team performed metagenomic sequencing on faecal samples from 1,627 children with or without ASD,aged one to 13 years old from five independent cohorts.Subjects were recruited from the Child and Adolescent Psychiatric Services of the Department of Psychiatry at Alice Ho Miu Ling Nethersole Hospital of the Hospital Authority’s New Territory East Cluster from 2021 to 2023.The team analysed faecal samples and clinical data including diet,medication and co-morbidities.The researchers identified apanel of novel gut microbiome markers including 14 archaea,51 bacteria,seven fungi,18 viruses,27 microbial genes and 12 metabolic pathways that were altered in children with ASD.Using machine learning approaches,they then developed anovel,non-invasive diagnostic model based on apanel of 31 multikingdom and functional markers that showed ahigh diagnostic accuracy for ASD.Dr Su Qi,Research Assistant Professor in the Department of Medicine and Therapeutics at CU Medicine,said,“Bacterial composition has been shown to be altered in ASD but the contribution of other components of the microbiome including the archaea,fungi,viruses,microbial genes or functional pathways remains unexplored.We found that the 31-microbiome panel has asensitivity of 94%and specificity of 93%for the diagnosis of ASD,and maintained asensitivity of 91%in children from an independent hospital cohort and ayounger community cohort from one to six years old.”Professor Siew Ng,Croucher Professor in Medical Sciences at CU Medicine,Director of the Microbiota I-Center(MagIC),and New Cornerstone Investigator added,“The diagnosis of ASD is challenging and requires regular developmental assessment in children who show signs of atypical social and language development.Diagnosis is often delayed especially in younger children who may only have mild symptoms and this could lead to delayed intervention.This,to our knowledge,is the first study to demonstrate the robustness and utility of anon-invasive biomarker to diagnose and predict ASD across different ages,gender and settings.” 查看详细>>

A model co-designed by aprofessor at Anglia Ruskin University is expected to reduce the need for chemotherapy in up to 38%of breast cancer patients who would previously have been advised to consider the treatment.The latest version of the PREDICT Breast model,published in the npj Nature journal and launched this week,uses the latest breast cancer survival data as well as taking into account the benefits and harms of chemotherapy and radiotherapy.PREDICT Breast was initially launched in 2010 by Gordon Wishart,Professor of Cancer Surgery at ARU and then Director of the Cambridge Breast Unit at Cambridge University Hospitals NHS Foundation Trust,and Paul Pharoah who at the time was Professor of Cancer Epidemiology at University of Cambridge.They brought together ateam of leading clinicians and scientists to develop and validate the PREDICT Breast model,which was based on Cancer Registry data from the UK.PREDICT Breast has been continuously updated since its launch and allows estimation of 10 and 15-year survival,as well as the absolute benefits of chemotherapy,trastuzumab,hormone therapy and bisphosphonates,to allow appropriate use of these therapies.The model is currently used worldwide by over 40,000 clinicians and their patients each month.The new version has been largely unfunded,but the recently published data is now being followed up by afurther study in the United States,using data from the SEER(Surveillance,Epidemiology,and End Results programme)database.Professor Wishart,now Chief Medical Officer at Check4Cancer alongside his visiting role at ARU,said:“Chemotherapy can cause significant physical effects such as nausea,weight loss,fatigue,bleeding,bruising and increased risk of infection.The data from the new model shows that for asignificant number of women with breast cancer,chemotherapy can be safely avoided. 查看详细>>

A newly discovered hormone that keeps the bones of breastfeeding women strong could also help bone fractures heal and treat osteoporosis in the broader population.Researchers at UC San Francisco and UC Davis showed that in mice,the hormone known as Maternal Brain Hormone(CCN3)increases bone density and strength.Their results,published July 10 in Nature,solve along-standing puzzle about how women’s bones remain relatively robust during breastfeeding,even as calcium is stripped from bones to support milk production.“One of the remarkable things about these findings is that if we hadn’t been studying female mice,which unfortunately is the norm in biomedical research,then we could have completely missed out on this finding,”said Holly Ingraham,PhD,the senior author of the new paper and aprofessor cellular molecular pharmacology at UCSF.“It underscores just how important it is to look at both male and female animals across the lifespan to get afull understanding of biology.”More than 200 million people worldwide suffer from osteoporosis,a severe weakening of the bones that can cause frequent fractures.Women are at particularly high risk of osteoporosis after menopause because of declining levels of the sex hormone estrogen,which normally promotes bone formation.Estrogen levels are also low during breastfeeding,yet osteoporosis and bone fractures are much rarer during this time,suggesting that something other than estrogen promotes bone growth. 查看详细>>

近日，中国科学院近代物理研究所的科研人员与来自法国、芬兰、南非和英国等国家的合作者首次成功测量了β缓发质子核镧-120的激发态结构，在质子滴线原子核的质子中子相互作用和形状演化的研究中取得重要进展，相关成果于近期发表在Physics Letters B上。理论预言，当位于中重质量区的原子核靠近N="Z线时，质子-中子相互作用会增强，并对激发态的结构产生重要影响。同时，原子核可能伴随形状的演化，呈现出“橄榄球”（长椭球），甚至是稀有的“南瓜形”（扁椭球）、“梨形”（八极形变）和“猕猴桃形”（三轴形变）。因此，通过实验测量奇特核的激发态性质对于检验相关理论模型至关重要。为了探索极端丰质子镧原子核的结构演化及其背后的物理机制，近代物理所和法国巴黎萨克雷大学的研究人员主导开展了寻找镧-120激发态的实验。镧-120是一种稀有的β缓发质子核，于1984年首次发现。由于熔合蒸发反应生成镧-120的截面极小，反应产物十分复杂，因此分离及鉴别镧-120极其困难。在过去的40年中，实验物理学家一直未能成功测量到镧-120的激发态。研究团队利用芬兰于韦斯屈莱大学重离子加速器上的质量分析谱仪和伽马探测器阵列，结合多种时间空间关联测量技术，首次在实验上建立了镧-120的激发态能级结构，发现镧-120的奇偶能级劈裂符合系统性，但是它的电磁跃迁比显著不同。结合理论模型，研究团队发现镧-120展现出一种稀有的三轴形变，并且质子-中子相互作用在描述质子滴线奇奇核的结构中扮演着重要角色。该研究得到了国家自然科学基金、中法科研伙伴交流计划项目和中国科学院未来伙伴网络专项的支持。 查看详细>>

Artificial intelligence(AI)could be used to identify drug resistant infections,significantly reducing the time taken for acorrect diagnosis,Cambridge researchers have shown.The team showed that an algorithm could be trained to identify drug-resistant bacteria correctly from microscopy images alone.Antimicrobial resistance is an increasing global health issue that means many infections are becoming difficult to treat,with fewer treatment options available.It even raises the spectre of some infections becoming untreatable in the near future.One of the challenges facing healthcare workers is the ability to distinguish rapidly between organisms that can be treated with first-line drugs and those that are resistant to treatment.Conventional testing can take several days,requiring bacteria to be cultured,tested against various antimicrobial treatments,and analysed by alaboratory technician or by machine.This delay often results in patients being treated with an inappropriate drug,which can lead to more serious outcomes and,potentially,further drive drug resistance.In research published in Nature Communications,a team led by researchers in Professor Stephen Baker’s Lab at the University of Cambridge developed amachine-learning tool capable of identifying from microscopy images Salmonella Typhimurium bacteria that are resistant to the first-line antibiotic ciprofloxacin–even without testing the bacteria against the drug.S.Typhimurium causes gastrointestinal illness and typhoid-like illness in severe cases,whose symptoms include fever,fatigue,headache,nausea,abdominal pain,and constipation or diarrhoea.In severe cases,it can be life threatening.While infections can be treated with antibiotics,the bacteria are becoming increasingly resistant to anumber of antibiotics,making treatment more complicated.The team used high-resolution microscopy to examine S.Typhimurium isolates exposed to increasing concentrations of ciprofloxacin and identified the five most important imaging features for distinguishing between resistant and susceptible isolates.They then trained and tested machine-learning algorithm to recognise these features using imaging data from 16 samples.The algorithm was able to correctly predict in each case whether or not bacteria were susceptible or resistant to ciprofloxacin without the need for the bacteria to be exposed to the drug.This was the case for isolates cultured for just six hours,compared to the usual 24 hours to culture asample in the presence of antibiotic. 查看详细>>

Everyday objects such as metal chains,handcuffs,and key rings are examples that demonstrate aunique combination of properties where hard,rigid rings are interlocked together to exhibit flexibility and strength as awhole,and as such enabling them to perfect their intended functions(Figure 1).At the molecular level,molecules composed of interlocked,nano-sized rings are known as catenanes,which are promising candidates for developing molecular switches and machines.Yet,due to their challenging synthesis,applications of catenanes in other areas are largely unexplored.Recently,a research team led by Professor Ho Yu AU-YEUNG from the Department of Chemistry at The University of Hong Kong(HKU)has synthesised acatenane composed of two freely-rotating rigid macrocycles and showed that the catenane can bind strongly and selectively to either copper(I)cation or sulfate anion despite their opposite charge and different geometry.The ability to detect and differentiate these specific ions has important implications for applications in areas like environmental monitoring and medical diagnostics.As same charges repel and opposite charges attract each other,a binding site that attracts apositively charged cation will normally experience arepulsive interaction with the negatively charged anion and vice versa,which made designing ahost that is suitable for both cation and anion very challenging.To overcome this challenge,the team installed both cation and anion binding sites on each of the interlocked rings,and by virtue of the rotatory motions of the catenane,the host can efficiently adjust the relative position of the binding sites and freely adapt aspecific form favourable for the spherical copper(I)cation or the tetrahedral sulfate anion,resembling achameleon that can change its appearance to fit in specific environments(Figure 2).This work has recently been published in the leading scientific journal Nature Communications.Apart from their industrial and environmental significance,both copper(I)and sulfate ion are essential for proper cell growth and organism development.The strong and selective binding to these ions by the catenane host could hence be exploited for the extraction and recycling of these ions from environmental samples.Also,just as the measurement of sodium ions,chloride and other electrolytes in blood samples can be aroutine test for blood pressure monitoring and general health,new technologies for selective recognition and binding of ions and minerals will be useful for diagnostic and therapeutic purposes.‘This work highlights catenane as an efficient candidate for potent molecular receptors with versatile structures,switchable properties and guest binding behaviours.’stated Professor Au-Yeung.In terms of future plans,Professor Au-Yeung and his group are developing more sophisticated catenane hosts for the simultaneous binding of multiple cations,anions and ion pairs. 查看详细>>

近日，清华大学交叉信息研究院邓东灵研究组与浙江大学物理学院王浩华、宋超研究组等合作，在超导系统中首次制备了斐波那契非阿贝尔拓扑态并实现了斐波那契任意子的编织操作。自然界中常见的基础粒子分为玻色子和费米子两种，交换两个基础粒子的位置会导致系统波函数产生+1（玻色子，如光子）或-1（费米子，如电子）的相位。这是由于在三维空间中，粒子A绕粒子B一圈（等价于交换位置两次）的环路可以在不经过粒子B的情况下连续变形至消失。这限制了系统在粒子交换两次后必须回到最初的量子态，因此每交换一次系统波函数只能产生+1或-1的相因子，相应的粒子被称为玻色子或费米子，满足玻色-爱因斯坦或费米-狄拉克统计规律。而在二维空间中，粒子A绕粒子B一圈的环路无法在不经过粒子B的情况下连续变形至消失，因此没有粒子交换两次后必须回到最初的量子态的限制。在此情形下，粒子的交换可以产生任意的相位，这样的粒子被称作阿贝尔任意子（Anyon），其交换位置的过程被称作编织（braiding）。更一般地，如果系统基态存在简并，交换两个粒子甚至可以改变系统波函数的振幅，导致系统整体的幺正演化而非仅获得一个全局相位。这种粒子被称为非阿贝尔任意子。非阿贝尔任意子的研究具有重要基础理论意义和潜在应用价值。此类粒子满足非阿贝尔统计规律，是与传统玻色子和费米子有着根本不同的奇异粒子。非阿贝尔任意子也是拓扑量子计算的基石。在拓扑量子计算中，量子门由非阿贝尔任意子的编织实现，计算结果的测量则由任意子的融合（fusion）完成。任意子的拓扑性质使得这种量子计算机天生对局域错误免疫，提供了硬件层面的容错量子计算方案。尽管存在多种理论方案，非阿贝尔任意子的实验实现十分困难，直到近年来才出现在量子处理器上模拟非阿贝尔任意子的工作。然而之前所有模拟的非阿贝尔任意子其编织操作所对应的量子门均不具备通用量子计算的能力。而斐波那契任意子则拥有更加复杂的统计性质，其实验实现更为困难。斐波那契任意子量子维度为黄金分割率1.618，与数学中的斐波那契序列息息相关（图1）。其编织能实现任意量子门，可以用于构建通用的容错量子计算机。实验制备斐波那契非阿贝尔拓扑态以及实现斐波那契任意子的编织操作被广泛认为极为困难。该实验采用弦网凝聚模型，通过几何变换使得超导量子芯片方形格子上的量子比特与弦网模型中蜂窝形状的“弦”相吻合（图2）。在该模型中，系统哈密顿量由所有涡旋算符Qv和所有块算符Bp之和构成，基态中所有的弦均为闭合，而激发态中斐波那契任意子分布在开弦的两端（图2）。该实验使用了27个超导量子比特，单（双）比特门精度为99.96%（99.5%），通过115层量子线路制备了系统基态。在制备基态之后，实验通过将系统分成不同区域的方法测量了拓扑纠缠熵，所得结果与理论预言吻合。在此基础上，实验通过弦算符操作产生了两对斐波那契任意子并展示了其编织操作（图3）。实验设计了多种不同的编织次序来测试斐波那契任意子的特性（图3a），分别为：（i）斐波那契任意子与其反粒子湮灭；（ii）编织改变融合结果；（iii）和（iv）融合结果相同验证Yang-Baxter方程；（v）测量斐波那契任意子的量子维度。实验所得结果均与理论预测吻合得很好（图3b），其中根据编制次序（v）的实验结果所得的斐波那契任意子量子维度为1.598，十分接近理论预言的黄金分割率1.618。作为拓扑量子计算领域重要的基础模型，斐波那契任意子的成功模拟与编织是实现通用拓扑量子计算的基础。该研究首次制备了斐波那契非阿贝尔拓扑态并实现了斐波那契任意子的编织操作，向最终实现通用拓扑量子计算迈出了重要一步。 查看详细>>

The current method for assessing medication-related liver injury is not providing an accurate picture of some medications’toxicity—or lack thereof—to the liver,according to anew study led by researchers from the Perelman School of Medicine.Classification of amedication’s potential to damage the liver,termed“hepatotoxicity,”has been historically determined by counting individual reported cases of acute liver injury(ALI).Instead,the researchers used real-world health care data to measure rates of ALI within apopulation and uncovered that some medications’levels of danger to the liver are being misclassified,in apaper published in JAMA Internal Medicine.“Incidence rates of severe ALI can be avaluable tool for determining amedication’s toxicity to the liver and when patients should be monitored,since incidence rates provide atruer,real-world look at this toxicity.Case reports did not accurately reflect observed rates of ALI because they do not consider the number of persons exposed to amedication,and cases of drug-induced liver injury are often underreported,”says senior author Vincent Lo Re.Within the study,17 different medications had rates that exceeded five severe ALI events per 10,000“person-years,”a measure that reflects both the amount of people in agroup and how long the study observes them.The team determined that 11 of these medications were in lower categories of hepatoxicity by case counts that were likely not reflective of their true risk,since their incidence rates revealed higher levels of toxicity.To determine incidence rates,Lo Re and his team,including lead author Jessie Torgersen,an assistant professor of medicine,examined electronic medical record data on almost 8million people.Each person did not have pre-existing liver or biliary disease when they began taking any of the 194 medications that were studied.Each of those medications were analyzed due to suspicion that they could cause harm to the liver,since each had more than four published reports of liver toxicity associated with their use.On the other side of the hepatotoxicity coin,the researchers found eight medications that were classified as the most hepatotoxic based on the number of published case reports,but should actually be in the least liver-toxic group,with incidence rates of less than one severe ALI event per 10,000 person-years.With these findings,the researchers hope that there might soon be mechanisms established within electronic medical records to alert clinicians to closely monitor the liver-related laboratory tests of patients who start amedication with ahigh observed rate of severe ALI. 查看详细>>

Sean Peters is leading amajor multi-institutional initiative to develop power efficient passive radar systems that could peek under the surface of Mars.Peters has earned a$2.45 million,three-year NASA grant to create adrone-based system to map subsurface areas.The project includes field-testing on Earth with an eye toward potential future deployments on missions to the red planet.The work will be carried out in collaboration with NASA’s Jet Propulsion Laboratory,the University of Arizona,and the Reykjavik University in Iceland.“This will allow us to understand the properties of the surface,the depth of ice deposits,and areas that have potential for astro-biological studies on indicators that may support life”said Peters,an assistant professor in the Ann and H.J.Smead Department of Aerospace Engineering Sciences at the University of Colorado Boulder.Utilizing radar on adrone presents unique challenges,and Peters’team has ideas how to solve this challenging problem.Most radar technology actively transmits signals,sending out pings and tracking the response to map nearby terrain or objects.This technology has been applied to various industries,such as military,air traffic control,and the geosciences.Aboard adrone,such systems are not always practical,as they are large and power hungry.Peters has proposed amuch smaller passive radar system that,instead of emitting its own signals,would pick up natural electromagnetic waves emitted by the sun and Jupiter to conduct measurements.“You’re listening for radio noise,essentially the unwanted part in atraditional active radar,to implement this low-resource technology onboard an uncrewed aerial system for altimetry and sounding,”Peters said.“We’ve done preliminary tests with the sun,and we know this is possible.It should be possible for Jupiter too.”Taking advantage of ambient radio waves from radio-astronomical bodies was afocus of Peters’PhD thesis and has been an area of active work for nearly adecade at NASA’s Jet Propulsion Laboratory,which is apartner on the grant.“Jupiter produces radio bursts at the same frequencies as traditional ground penetrating radars,and we can measure them here on Earth.They penetrate into the ground,and our goal is to pick up and analyze the reflected signals to observe what’s below the surface,”Peters said.Peters’PhD student,Thorsteinn Kristinsson,is conducting early work on the grant.“We feel electromagnetic waves coming from the sun just going outside.You wouldn’t think looking at the sky that there are waves hitting your body from Jupiter too,but at certain times there are.”The project is incorporating both the sun and Jupiter because their electromagnetic waves cover different areas of the frequency spectrum.Jupiter’s waves are lower frequency and penetrate deeper into the ground,allowing the team to conduct additional subsurface analysis.The team will design and build the radar system and conduct initial field-testing on the ground in California within the next year.By the third year of the grant,the radar system will be incorporated into adrone for flight tests in Iceland,which has terrain analogous to Martian volcanoes.The Iceland portion of the grant is particularly exciting for Kristinsson,who grew up there and has conducted previous research in the same area.“It’s amazing.It gives me the opportunity to do work in my home country and field testing in an environment Iknow,and that is also so beautiful to be in,”Kristinsson said. 查看详细>>

Artificial intelligence models often play arole in medical diagnoses,especially when it comes to analyzing images such as X-rays.However,studies have found that these models don’t always perform well across all demographic groups,usually faring worse on women and people of color.These models have also been shown to develop some surprising abilities.In 2022,MIT researchers reported that AI models can make accurate predictions about apatient’s race from their chest X-rays—something that the most skilled radiologists can’t do.That research team has now found that the models that are most accurate at making demographic predictions also show the biggest“fairness gaps”—that is,discrepancies in their ability to accurately diagnose images of people of different races or genders.The findings suggest that these models may be using“demographic shortcuts”when making their diagnostic evaluations,which lead to incorrect results for women,Black people,and other groups,the researchers say.“It’s well-established that high-capacity machine-learning models are good predictors of human demographics such as self-reported race or sex or age.This paper re-demonstrates that capacity,and then links that capacity to the lack of performance across different groups,which has never been done,”says Marzyeh Ghassemi,an MIT associate professor of electrical engineering and computer science,a member of MIT’s Institute for Medical Engineering and Science,and the senior author of the study.The researchers also found that they could retrain the models in away that improves their fairness.However,their approached to“debiasing”worked best when the models were tested on the same types of patients they were trained on,such as patients from the same hospital.When these models were applied to patients from different hospitals,the fairness gaps reappeared.“I think the main takeaways are,first,you should thoroughly evaluate any external models on your own data because any fairness guarantees that model developers provide on their training data may not transfer to your population.Second,whenever sufficient data is available,you should train models on your own data,”says Haoran Zhang,an MIT graduate student and one of the lead authors of the new paper.MIT graduate student Yuzhe Yang is also alead author of the paper,which appears today in Nature Medicine.Judy Gichoya,an associate professor of radiology and imaging sciences at Emory University School of Medicine,and Dina Katabi,the Thuan and Nicole Pham Professor of Electrical Engineering and Computer Science at MIT,are also authors of the paper. 查看详细>>