Blog

  • Something like feathers grew on a 247-million-year-old reptile

    Something like feathers grew on a 247-million-year-old reptile

    German paleontologists have discovered a 247-million-year-old fossil of a reptile with a bizarre row of plumes sprouting from its back. The elaborate display is a paradox of evolution. The plumes bear some similarities to feathers, even though the newly discovered reptile was not closely related to birds.

    Stephan Spiekman, a paleontologist at the Stuttgart State Museum of Natural History in Germany and an author of the new study, said that the discovery could change how scientists think about the origin of feathers. In birds, a complex network of genes is enlisted to sprout feathers from their skin. Part of the network might have already evolved in early reptiles more than 300 million years ago.

    If that’s true, Spiekman said, it would mean that other ancient reptiles might have sprouted strange ornaments of their own that are waiting to be discovered.

    “I hope this will broaden our perspective,” Spiekman said. “And then who knows what we’ll find?”

    In their study, which was published Wednesday in the journal Nature, Spiekman and his colleagues named the reptile Mirasaura grauvogeli. In Latin, Mirasaura means “wonderful reptile.” And grauvogeli honors Louis Grauvogel, the French paleontologist who dug up the fossil in 1939.

    Grauvogel was a wealthy factory owner with training in biology. He spent much of his free time looking for fossils in the quarries of northeastern France, and by the time he died in 1987, he had built up a huge private collection of animal and plant remains. His daughter, Lea Grauvogel-Stamm, herself an accomplished paleontologist, donated the fossils to the Stuttgart Museum in 2019.

    When Grauvogel first uncovered Mirasaura in 1939, he could see only the animal’s crest exposed in a rock. He speculated that he had found the fin of a fish.

    Eighty years later, when the Stuttgart scientists began inspecting Grauvogel’s collection, they noticed that his supposed fish fin was actually connected to a reptile bone at one end. The rest of the bone was hidden in the rock.

    The researchers picked away at the rock and discovered the rest of Mirasaura’s skeleton. A further inspection of Grauvogel’s fossil collection revealed more crests, along with a second skeleton.

    Looking at Mirasaura, Spiekman was immediately reminded of one of the most mysterious reptile fossils ever found, a 220-million-year-old creature called Longisquama insignis.

    Discovered in Central Asia in 1969, Longisquama’s fossil preserved impressions of long, flat projections extending from its back. Its discoverers speculated that these were elongated scales that had fanned out to either side of Longisquama’s body. The reptile used them like parachutes, they claimed, slowing its fall as it jumped from trees.

    In 2000, a team of American researchers offered a controversial new theory: Longisquama’s parachute scales were actually feathers, and Longisquama might be an ancient relative to today’s birds.

    That view eventually fell out of favor, as paleontologists subsequently discovered many dinosaurs with feathers dating to 160 million years ago. Some of these structures were almost as complex as bird feathers; others were simple wires. It is now clear that birds are living dinosaurs.

    Longisquama drifted into a scientific limbo: No one could say what kind of reptile it was, and the nature of its plumes remained anyone’s guess. “The consensus became, ‘We really don’t know what Longisquama is – it’s a weird reptile,’” Spiekman said.

    With the discovery of Mirasaura, Longisquama gains a cousin. And even though Mirasaura lived almost 30 million years earlier than Longisquama, its fossils were in far better shape. Spiekman and his colleagues could study its crest in microscopic detail and inspect its exquisitely preserved skull.

    Their analysis shows that Mirasaura and Longisquama belonged to an extinct lineage of reptiles that specialized in living in trees. That lineage is only distantly related to birds and dinosaurs, having split off on its own more than 300 million years ago.

    Based on that finding, the scientists argue that the plumes of Mirasaura and Longisquama evolved from ordinary reptile skin. Birdlike dinosaurs independently evolved feathers.

    A close inspection of Mirasaura’s crest supported that conclusion, revealing some fundamental differences from feathers. Feathers are made of branching fibers, for example, while Mirasaura sported stiff sheets that grew from a central ridge.

    But Spiekman and his colleagues also concluded that Mirasaura’s crest bore some important similarities to feathers. Feathers gain some of their color from microscopic sacs of pigment called melanosomes. Mirasaura’s crest also contains melanosomes, and they have the same shape as feather melanosomes.

    The shape of Mirasaura’s plumes also suggests to Spiekman that they grew in a featherlike way, developing from a ring of cells that rose up from the skin before fanning out.

    If the researchers are right, then the common ancestor of Mirasaura and birds must have already carried some of the genetic instructions for building featherlike growths. Only some reptiles went on to use those instructions to that end.

    As for how Mirasaura used its crest, Spiekman discounts the idea of parachuting from the trees. The new fossils clearly show that the crest stood straight up on the animal’s back – a position not conducive to slowing a fall.

    “And that, for us, only really leaves some sort of display as a potential option,” Spiekman said. One possibility is that Mirasaura used its crest to show off, akin to how male anole lizards attract mates with a colorful flap of skin under their jaw.

    These ideas are likely to set off a new round of debates. Richard Prum, an ornithologist at Yale University who was not involved in the new research, questioned whether Mirasaura’s crest had much in common with feathers. “I think that’s a real mistake,” he said.

    He also held out the possibility that the crest was not a sheet of dead cells but, perhaps, a sheet of living tissue with blood coursing through it. “Thermal regulation becomes a prospect – absorbing the sun so they can get lots of energy to run around quicker,” Prum speculated.

    Whatever Mirasaura’s strange anatomy turns out to have been, Prum agreed that the find highlights the underappreciated versatility of reptile skin.

    “It’s a fascinating thing,” he said.


    This article originally appeared in The New York Times.


    Continue Reading

  • Four more traders appeal rate-rigging convictions after Supreme Court ruling

    Four more traders appeal rate-rigging convictions after Supreme Court ruling

    Four traders are appealing to have their rate-rigging convictions overturned after the Supreme Court quashed two rate-rigging cases on Wednesday.

    Jay Merchant, Jonathan Mathew, Philippe Moryoussef, and Christian Bittar are seeking acquittal following the victory of traders Tom Hayes and Carlo Palombo.

    All of the traders were convicted of manipulating the interest rates used for loans between banks, know as Libor in the UK, an issue at the heart of the 2008 financial crisis.

    “Following the Supreme Court’s landmark decision yesterday to quash the convictions of Tom Hayes and Carlo Palombo, all four of our clients now intend to appeal against their convictions,” said law firm Hickman & Rose.

    “In those circumstances, they don’t intend to comment further at this time,” the firm added.

    The four convictions came after an investigation from the Serious Fraud Office into whether traders had been manipulating Libor for profit.

    Libor became the focus of allegations of wrongdoing following the financial crisis in 2008 and has now been discontinued, while its European equivalent Euribor is being reformed.

    Because the Supreme Court has now ruled in favour of Mr Hayes and Mr Palombo, the four traders’ appeal is likely to be a more straightforward process than for Mr Hayes and Mr Palombo who argued their case for years.

    The Serious Fraud Office declined to comment on the appeal from the four traders on Thursday.

    However, it said on Wednesday in response to the ruling on Mr Hayes and Mr Palombo’s case that it had “considered this judgement and the full circumstances carefully and determined it would not be in the public interest for us to seek a retrial”.

    The Libor scandal came to light in 2012, when it was discovered that banks were artificially inflating rates to profit from trading and were also lowering them to mask the troubles they faced following the outbreak of the global financial crisis.

    However, in 2023, the BBC uncovered evidence of a much larger, state-led “rigging” of interest rates, under pressure from central banks and governments across the world during the financial crisis.

    Mr Hayes and Mr Palombo argued they were wrongly prosecuted for what were normal commercial practices in order to appease public anger towards the banks over the financial crisis.

    Continue Reading

  • How Rust-Based Zed Built World’s Fastest AI Code Editor

    How Rust-Based Zed Built World’s Fastest AI Code Editor

    When Nathan Sobo set out to build the Zed code editor, he had a simple but ambitious goal: Make it fast and fun.

    Zed Industries (Zed, for short) is building what it expects to be the world’s fastest text editor for hyper-responsive and collaborative coding experiences. It offers tools for native debugging support in multiple languages, agentic editing and edit prediction, with more features planned on the roadmap.

    The Zed code editor and its AI capabilities are all open source, including Zeta, the open large language model (LLM) that powers Zed’s edit prediction.

    Performance Is Key

    “You hit a key or interact with the app in any way, and we have pixels available for you with zero perceptible lag,” Sobo told The New Stack. “Zero perceptible lag means we need it on the next refresh of your monitor.”

    That kind of performance seemed impossible in an era where most code editors are built on web technologies. But Sobo, who previously led GitHub‘s Atom project, knew exactly why that approach would not work. “You’re not going to do that with JavaScript ever,” he said bluntly.

    Yet, three years after founding Zed Industries in 2021, the company has built a Rust-powered IDE that’s attracting tens of thousands of developers with its combination of native performance and real-time collaboration features. But as AI capabilities became “table stakes” for developer tools, Zed faced a new challenge in how to deliver AI-powered code completion that matched its zero-latency philosophy.

    The solution came through a partnership that transformed its inference infrastructure in just one week, Sobo said.

    The AI Integration Challenge

    By 2024, Zed had built a solid foundation for the company’s performance-first editor, but it knew AI features were becoming essential. “New features coming from language models and the revolution in language models are becoming pretty much table stakes and expected as part of the experience of what it means to write software,” Sobo told The New Stack in an insightful exchange.

    The company developed “edit prediction” — an AI feature powered by its Zeta model that anticipates what developers want to do next and suggests changes in real time. Built on the Qwen series of open source models, the system uses a technique called speculative decoding to optimize for the specific case of code editing.

    “The idea of an edit implies most of the text is unchanged, but a few parts might be changed,” Sobo explained. “What speculative decoding does is it says every time we get a run of tokens out of the model that matches something that’s in the input, why don’t we just assume that we’re going to keep matching it a little bit longer, and run ahead.”

    However, with its launch set for just a week away, Zed’s existing inference provider was not delivering the performance the company needed, Sobo said.

    They were missing critical targets: P90 latency under 500ms and P50 under 200ms. Even worse, the provider offered limited compute capacity, no multiregion support, and what Sobo calls a “black box” approach that conflicted with Zed’s open source values.

    “They were very kind of turnkey in nature,” Sobo said.

    With open source as one of its core values, Zed wanted more visibility into what was driving its model performance so the team could grow its own expertise and find iterative improvements, he noted.

    Engineering-First Partnership

    Enter Baseten, an inference platform provider that took a different approach. Instead of a hands-off service, it deployed forward engineers directly to Zed’s problem. Within days, the team had tested over 75 different performance optimizations.

    “Baseten showed up with an outstanding level of engagement,” Sobo said. “They led with engineering and led with competence. … Watching them traverse the curve from where they were, which is like, ‘Hey, I love you guys as people, I love how you’re showing up, but the number isn’t good,’ to ‘trust us, we will make this number, we’ll move it where it needs to be in time for your deadline.’ And then seeing them do that, that’s pretty dope.”

    The technical solution included the TensorRT-LLM framework instead of vLLM as the inference framework; KV caching and custom-tuned speculative decoding to massively reduce latency; lookahead decoding for higher throughput; and multicloud capacity management with geo-aware routing. The team also custom-tuned autoscaling settings to ensure optimal resource utilization while maintaining low latency.

    “Getting their direct, hands-on help actually getting this model that we developed deployed doing speculative decoding was super key,” Sobo noted. “We’re very CPU people. When we touch GPU, it’s to run graphics shaders, right? Getting this model running on hardware — that was not our core competency at all.”

    The Results: 2x Performance Improvement

    The partnership delivered results that exceeded Zed’s initial goals:

    • 45% lower P90 latency
    • 3.6x higher throughput
    • 100% uptime
    • Over 2x faster edit prediction compared to its previous provider

    In addition, the migration was seamless, Sobo said. Because Baseten maintained OpenAI compatibility, Zed moved all its traffic over within a single day with no code changes required.

    Moreover, the performance improvements did not stop at launch. Baseten’s team continued iterating, eventually shipping a custom “Baseten Lookahead” decoding method that shaved hundreds of additional milliseconds off prediction times.

    Beyond Performance: A Philosophy Match

    For Sobo, the technical achievements were only part of the story. The partnership was also a lesson in how companies should work together.

    “You have this finite amount of time on this planet. And who do I want to actually spend that time interacting with?” he said. “These guys just showed up in a way that made me feel like I like these guys, like I actually want to do business with them. And that was the overriding concern.”

    This philosophy extends to how Sobo thinks about the entire developer tools landscape. While Zed competes directly with VS Code for individual developers, Sobo has even bigger ambitions: transforming how software teams collaborate.

    “Git is literally a tool that Linus Torvalds designed to manage patches that were being mailed to the Linux kernel mailing list,” he pointed out. “It’s literally a tool designed around email.”

    Zed’s vision involves real-time, fine-grained collaboration that goes far beyond traditional commit-based workflow, Sobo said. For instance, imagine highlighting code and instantly connecting with whoever wrote it or having persistent conversations that survive code changes and refactoring, he added.

    “So, the vision is long-term to eat into the software collaboration market, but the beachhead we’re trying to get is just making developers love using our tool once we’ve claimed that real estate,” he added.

    The Rust Foundation

    None of this would be possible without Zed’s foundational choice to build in Rust. When Sobo started the project in 2018-2019, he saw no alternative for the kind of system-level performance he demanded.

    “You need a systems programming language to implement something of the complexity of an editor if you want it to actually be responsive,” he explained.

    “So, with Zed, after the experience with Atom and building on web technology, building a web page masquerading as a desktop app — which is what all these Electron apps are — I knew that web tech was never going to get me to the level of responsiveness that I really wanted.

    “You’re not going to do that with JavaScript, ever. Like, you need the control over the memory, multithreading, shared memory across multiple threads, and what Rust offers is the ability to do all of that while maintaining a lot of the productivity advantages of some of these more managed languages, like JavaScript — slower to develop than TypeScript, I’m not going to lie, but the result runs at a speed that would not be obtainable. And it’s definitely faster to develop Rust than C++ or C for us, it rules out large categories of things that could go wrong statically.”

    Zed has an open source code base with lots of people contributing “all over the place,” Sobo said.

    And “Doing that in a world where you can implement a use-after-free or there’s no compiler support for avoiding some of these very difficult-to-debug, potentially very dangerous [bugs] … would have rendered Zed, with a small team, impossible to build, whereas Rust made it possible.”

    AI Without Intellectual Laziness

    As AI tools become more powerful, Sobo has developed views on how to use them effectively. He said he is a heavy user of AI for coding, often switching between models for different tasks — using lighter models for codebase exploration, then switching to more powerful models for complex implementation work.

    But he’s wary of what he calls “intellectual laziness.”

    Said Sobo: “If you could literally enter three prompts and do the dishes while this thing’s vomiting this thing out, is it that valuable? Like, do you even want this? Or what can you learn from this and then literally throw it in the trash and start again.”

    This philosophy reflects his approach to building Zed itself — use every tool available but never lose sight of the goal of building better systems.

    What’s Next?

    Meanwhile, with its AI infrastructure optimized and a growing user base, Zed faces the challenge of any developer tool: feature completeness. The editor still lacks Windows support, which has generated constant requests from users. And the company is working to expand the tool’s debugging capabilities and other IDE essentials, he said.

    Still, Sobo said he remains focused on the longer-term vision of transforming developer collaboration. The company’s recent work has revived early collaboration features, with plans for persistent code permalinks, real-time shared editing and what Sobo calls “metadata layers” on top of code that can capture human and AI conversations.

    “There’s this sense of, like, the code is where we hang out together, kind of like people that are friends on ‘World of Warcraft,’ like they hang out together in their whatever ‘World of Warcraft’ universe,” Sobo said. “I’ve never played it. For me, my ‘World of Warcraft’ universe is the Zed code base, where I hang out with my friends and talk about the code, talk about the changes,” he said, describing how Zed’s distributed team already works. The goal is to scale that experience to the broader developer community.

    The Baseten partnership proved that when it comes to performance-critical infrastructure, engineering-first relationships can deliver transformative results.


    Group Created with Sketch.


    Continue Reading

  • Renesas Announces Loss Resulting from Signing Restructuring Support Agreement with Wolfspeed

    Renesas Announces Loss Resulting from Signing Restructuring Support Agreement with Wolfspeed

    TOKYO, Japan ― Renesas Electronics Corporation (TSE: 6723, “Renesas”), a premier supplier of advanced semiconductor solutions, today announced the finalized amount of the loss, previously disclosed as an estimate in the announcement titled “Renesas Announces Expected Loss Resulting from Signing Restructuring Support Agreement with Wolfspeed” on June 23, 2025.

    1. Recording of Loss

    Renesas previously announced that there was a possibility of recording a loss of approximately 250 billion yen (converted at an average exchange rate of 150 yen to the dollar during the period) on the deposited receivables related to a deposit provided to Wolfspeed, Inc. (NYSE: WOLF, “Wolfspeed”) in its consolidated financial statements for the six months ended June 30, 2025. Following a thorough review based on the restructuring plan that Wolfspeed filed with the U.S. court under Chapter 11 of the U.S. Bankruptcy Code and other related matters, Renesas recorded a loss of 235 billion yen in the consolidated financial results for the six months ended June 30, 2025. 

    2. Future Outlook

    The loss mentioned above has been reflected in the “Renesas Reports Financial Results for the Second Quarter Ended June 30, 2025” announced on July 25, 2025.

     

    About Renesas Electronics Corporation

    Renesas Electronics Corporation (TSE: 6723) empowers a safer, smarter and more sustainable future where technology helps make our lives easier. A leading global provider of microcontrollers, Renesas combines our expertise in embedded processing, analog, power and connectivity to deliver complete semiconductor solutions. These Winning Combinations accelerate time to market for automotive, industrial, infrastructure and IoT applications, enabling billions of connected, intelligent devices that enhance the way people work and live. Learn more at renesas.com. Follow us on LinkedInFacebookTwitterYouTube, and Instagram.


    Continue Reading

  • The Best Time to Exercise If You’re Trying to Lower Your Cholesterol

    The Best Time to Exercise If You’re Trying to Lower Your Cholesterol

    • Exercise can improve cholesterol levels, no matter when you do it.
    • Morning, afternoon or evening—consistency matters more than timing.
    • For long-term health benefits, include both cardio and strength training in your routine.

    Getting lab results back from your doctor can be nerve-wracking—especially if your cholesterol levels come back higher than expected. But it might be the wake-up call you need to make some lifestyle changes, like adding exercise back in your routine, to help bring down those numbers.

    Exercise has been shown to help improve cholesterol by lowering “bad” LDL cholesterol and raising “good” HDL cholesterol. But if you’re looking to get the most out of your efforts, you might wonder: Is there a best time of day to work out to help lower my cholesterol levels?

    We asked a cardiologist, sports dietitian and personal trainer—and the truth is, there isn’t a one-size-fits-all answer. When it comes to the best time to exercise for lower cholesterol, it depends. Keep reading to find out why. 

    Is There a Best Time?

    There isn’t a clear consensus on whether the morning, afternoon or evening is better for lowering cholesterol.

    “It depends on individual factors, but there’s emerging evidence that morning exercise may have a slight edge for cholesterol management,” says Aaron Feingold, M.D., a board-certified cardiologist. He explains that the body’s natural circadian rhythm—also known as your 24-hour biological clock—affects fat metabolism, with cholesterol production peaking at night. By exercising in the morning, you can help interrupt this cycle, leading to improved “good” HDL cholesterol levels throughout the day.

    For instance, one study on older adults found that morning exercisers had significantly lower “bad” LDL cholesterol levels than those who exercised in the afternoon. Another study reported that morning workouts were linked with an overall lower risk for coronary artery disease and stroke—two conditions where cholesterol plays a major role. 

    However, afternoon or evening exercise sessions have their own advantages. “Research suggests exercising in the late afternoon or evening may be more effective than morning exercise for lowering blood cholesterol,” says Marie Spano, M.S., RD, CSSD, CSCS, a sports dietitian and strength and conditioning specialist. She explains that some studies suggest hormonal fluctuations, enzyme activity and increased insulin sensitivity later in the day may increase fat clearance from the blood after exercise, potentially lowering cholesterol levels. 

    For example, one study found that exercising in the afternoon or evening led to a 25% reduction in insulin resistance—a condition linked to high cholesterol—indicating a potential benefit for blood lipid levels.

    Clearly, exercising at any time of day offers health benefits. Ultimately, experts agree that the best time to exercise is whatever works best for you. “The most important factor is consistency—the best time is whatever time allows someone to exercise regularly,” says Feingold. 

    Spano agrees and adds, “Exercise is beneficial regardless of the time of day one chooses to work out. If you’re more likely to exercise if you do it first thing in the morning, then that is the best time for you.”

    Benefits of Exercising in the Morning

    Lacing up your sneakers early in the day may offer several benefits. “When you move first thing in the morning, there’s less room for life’s daily obstacles and your own excuses to get in the way,” says Amanda Katz, CPT, a NASM-certified personal trainer and RRCA-certified running coach. She adds that morning workouts encourage you to make healthier choices the rest of the day, such as opting for a cholesterol-lowering breakfast like oats and berries instead of a pastry. 

    “Morning workouts can also improve insulin sensitivity throughout the day, leading to better glucose and lipid handling,” adds Feingold. Improving insulin sensitivity not only helps regulate blood sugar levels but also promotes healthier cholesterol levels.

    Morning exercise also capitalizes on naturally higher testosterone and growth hormone levels, which can enhance fat metabolism and muscle protein synthesis—factors that support better body composition and may help to lower cholesterol levels, explains Feingold.

    Benefits of Exercising in the Afternoon (or Evening)

    After a long day, exercise can be the perfect stress-reliever. “Working out in the afternoon or evening can lower stress levels and promote relaxation, both of which contribute to improved sleep quality,” says Katz. Ongoing stress and lack of sleep causes a sustained increase in levels of cortisol—a stress hormone—which can lead to dyslipidemia, an unhealthy imbalance of blood fats. By supporting your sleep and reducing stress levels, this could benefit your cholesterol levels.

    Working out in the evening may also improve your performance, leading to health benefits. “Core body temperature is typically higher in the afternoon, which can improve exercise performance and potentially lead to greater caloric expenditure. Perceived exertion is often lower in the afternoon, allowing for more intense workouts that can boost HDL cholesterol more effectively,” says Feingold.

    Getting Started

    If you’re feeling inspired by all the ways that exercise can help lower your cholesterol, here are some expert-backed tips to get you started. 

    • Break Up Activity into Manageable Chunks. Experts recommend 150 minutes of moderate-intensity exercise per week, which can be as simple as 30 minutes a day, five days a week. You can do it all at once or break it up into shorter sessions throughout the day.  
    • Be Intentional. Plan your workout ahead of time and be realistic. “Can you commit to two group fitness classes or personal training sessions a week?” says Katz. Putting it on your calendar makes it more likely to happen. 
    • Pick Activities You Enjoy. Choose a workout you will stick with and enjoy, says Spano. Whether it’s walking, swimming, dancing or cycling, consistency matters more than the type of activity, especially when you’re first starting out. 
    • Do Both Aerobic and Resistance Training. Feingold recommends doing both aerobic and resistance training each week, as each plays a role in overall fitness and healthy cholesterol levels.
    • Make Yourself Accountable. Hire a personal trainer or work out with a friend, recommends Spano. “You’re less likely to skip exercise sessions if you paid for it or somebody is counting on you to show up.”
    • Remember Your ‘Why.’ Remind yourself why you are starting to exercise. While lowering your cholesterol may be your initial goal, it can also boost energy, build muscle and support long-term health and longevity. 

    Cholesterol-Lowering Meal Plan To Try

    You Just Found Out You Have High Cholesterol—Try This 30-Day Meal Plan, Created by a Dietitian

    Our Expert Take 

    The best time to exercise for lowering your cholesterol is one that you can stick to consistently. This can look different from day to day or week to week depending on your schedule. Fortunately, research supports working out in the morning, afternoon and evening, meaning you’ll see benefits regardless of your workout time. If you’re just getting started with exercise, remember that consistency is key for long-term results.

    Continue Reading

  • NASA, SpaceX target July 31 for next crewed mission to International Space Station-Xinhua

    LOS ANGELES, July 24 (Xinhua) — NASA and SpaceX are targeting July 31 for the launch of their next crewed mission to the International Space Station (ISS), the U.S. space agency announced Thursday.

    Liftoff is scheduled for 12:09 p.m. U.S. Eastern Time (1609 GMT) next Thursday from NASA’s Kennedy Space Center in Florida. Docking with the ISS is expected around 3 a.m. on Saturday, Aug. 2.

    The new mission, codenamed Crew-11, will send NASA astronauts Zena Cardman and Mike Fincke, Japan Aerospace Exploration Agency astronaut Kimiya Yui, and Roscosmos cosmonaut Oleg Platonov to the ISS.

    This will mark the 11th crew rotation mission conducted in partnership with SpaceX under NASA’s Commercial Crew Program. The initiative aims to provide safe, reliable and cost-effective transportation to and from the ISS using commercial spacecraft.

    Continue Reading

  • Intel cuts back spending, workforce as struggling chip maker mounts comeback

    Intel cuts back spending, workforce as struggling chip maker mounts comeback

    Intel Corp. is shedding thousands of workers and cutting expenses as its new CEO works to revive the fortunes of the struggling chipmaker that helped launch Silicon Valley but has fallen behind rivals like Nvidia Corp.

    In a memo to employees, CEO Lip-Bu Tan said Intel plans to end the year with 75,000 workers, down 31% from 108,900 employees at the end of last year, through layoffs and attrition. The company previously announced a 15% workforce reduction.

    “I know the past few months have not been easy. We are making hard but necessary decisions to streamline the organization, drive greater efficiency and increase accountability at every level of the company,” Tan wrote.

    In addition, Intel will scrap previously planned projects in Germany and Poland and also move assembly and test operations in Costa Rica to larger sites in Vietnam and Malaysia. Costa Rica will remain a “home to key engineering teams and corporate functions,” Tan said in the memo.

    In the U.S., the company said it will “further” slow construction of a semiconductor plant in Ohio.

    Founded in 1968 at the start of the PC revolution, Intel missed the technological shift to mobile computing triggered by Apple’s 2007 release of the iPhone, and it’s lagged more nimble chipmakers. Intel’s troubles have been magnified since the advent of artificial intelligence — a booming field where the chips made by once-smaller rival Nvidia have become tech’s hottest commodity.

    The Santa Clara, California-based company’s market cap was $98.71 billion as of the market close on Thursday, compared with Nvidia’s $4.24 trillion.

    Continue Reading

  • Asia stock market live updates

    Asia stock market live updates

    Asia markets start Friday trading in the red

    Asia markets started the trading day lower.

    Japan’s benchmark Nikkei 225 fell 0.24%, while the Topix lost 0.55%.

    South Korea’s Kospi was flat and the small-cap Kosdaq slipped 0.48%.

    Australia’s S&P/ASX 200 lost 0.41%.

    —Lee Ying Shan

    Asia markets set to open lower

    Good morning from Singapore. Asia markets were set for a muted open Friday.

    Japan’s benchmark Nikkei 225 was set to start the trading day lower, with the futures contract in Chicago at 41,740 and its counterpart in Osaka at 41,640, against the index’s last close at 41,826.34.

    Futures for Hong Kong’s Hang Seng index stood at 25,505, pointing to a weaker open compared with the HSI’s last close of 25,667.18.

    Australia’s S&P/ASX 200 was also on track to start the day lower with futures tied to the benchmark at 8,637, compared with its last close of 8,709.4.

    — Lee Ying Shan

    S&P 500, Nasdaq close higher Thursday

    The S&P 500 and the Nasdaq Composite ended Thursday’s session with fresh record closes after scoring new all-time intraday highs earlier in the trading day.

    The broad market index closed 0.07% higher to end at 6,363.35, while the tech-heavy Nasdaq inched up 0.18% to finish at 21,057.96.

    However, the blue-chip Dow Jones Industrial Average pulled back 316.38 points, or 0.7%, to settle at 44,693.91.

    — Sean Conlon

    Continue Reading

  • Scientists open new atlas of genetic diversity with advanced sequencing

    Scientists open new atlas of genetic diversity with advanced sequencing

    A landmark study harnesses long-read sequencing to reveal vast, previously undetected structural variations in human DNA, reshaping our understanding of genetics and disease potential.

    Study: Structural variation in 1,019 diverse humans based on long-read sequencing

    In a recent study published in the journal Nature, researchers investigated large-scale structural variants (SVs), complex and poorly understood insertions, deletions, and rearrangements in DNA, using next-generation ‘long-read’ sequencing. Their groundbreaking dataset comprised 1,019 individuals across 26 global populations. The study further leveraged a novel graph-based analytical framework, allowing for the creation of over 107,000 sequence-resolved biallelic SVs, which the authors made open-access.

    The high-resolution genomic investigation not only significantly furthers our understanding of the true diversity of human genetics but also progresses our identification and future management of disease-causing genetic variants in patients.

    Background

    Biology textbooks often depict the human genome as a linear string of three billion combinations of A, T, G, and C – our DNA, the building blocks of our lives. The reality, however, is far more dynamic, with our DNA demonstrating large-scale structural variants (SVs)—deletions, duplications, insertions, and inversions of entire DNA segments.

    Despite accounting for most base-pair (bp) differences between any two organisms and being major contributors to and modulators of human health, they remain notoriously difficult to study and poorly understood. Short-read sequencing, the predominant sequencing technology of today, splices long DNA segments into tiny fragments, which are then amplified. While effective for small variants, these technologies struggle to map complex SVs, especially large insertions and multiallelic variable number tandem repeats (VNTRs), which are sometimes missed entirely.

    Consequently, a vast majority of the human genome remains invisible to science and medicine, allowing potentially curable genetic diseases to persist unabated. Long-read sequencing is a relatively novel technology that can read much longer, continuous stretches of DNA, thereby overcoming short-read sequencing’s primary SV-associated shortcoming. Harnessing this technology could unlock this hidden portion of the human genome and the medical treasures that lie within.

    About the study

    The present work does just this: A consortium of researchers undertook a massive, multinational project to map SVs using a globally diverse cohort. Study samples were acquired from the 1000 Genomes Project (1kGP) and initially comprised 1,064 samples (lymphoblastoid cell lines).

    Strict quality control (QC) using a combination of DNA concentration determination (multimode microplate reader), DNA purity evaluation (spectrophotometer), and DNA fragment length verification (Femto Pulse system) reduced the dataset to 1,019. This dataset comprised participants from 26 distinct ancestries across Africa, the Americas, Europe, and East and South Asia.

    a, Breakdown of self-identified geographical ancestries for 1,019 long-read genomes representing 26 geographies (that is, populations) from 5 continental regions. The three-letter codes used are equivalent to those used in the 1kGP phase III18 and are resolved in Supplementary Table 2. b, ONT sequence coverage per sample, expressed as fold-coverage (left), and N50 read length in base pairs (right). c, Schematic of the SAGA framework for graph-aware discovery and genotyping of SVs using a pangenome graph augmentation approach. Basemap in a from Natural Earth data (https://www.naturalearthdata.com).a, Breakdown of self-identified geographical ancestries for 1,019 long-read genomes representing 26 geographies (that is, populations) from 5 continental regions. The three-letter codes used are equivalent to those used in the 1kGP phase III18 and are resolved in Supplementary Table 2b, ONT sequence coverage per sample, expressed as fold-coverage (left), and N50 read length in base pairs (right). c, Schematic of the SAGA framework for graph-aware discovery and genotyping of SVs using a pangenome graph augmentation approach. Basemap in a from Natural Earth data (https://www.naturalearthdata.com).

    The long-read sequencing platform used was the Oxford Nanopore Technologies (ONT) LRS, a cutting-edge technology capable of generating data with a median read length of over 20,000 base pairs.

    To analyze this complex dataset, they engineered a novel computational framework called SAGA (SV analysis by graph augmentation). This process involved four key steps: First, aligning long reads to both linear (GRCh38) and graph-based (HPRC) references; second, SV discovery using Sniffles, DELLY, and the graph-aware SVarp algorithm, including specialized remapping to resolve inversion alignment artifacts; third, augmenting the pangenome graph to incorporate new SVs despite complexities in multiallelic VNTR genotyping; and finally, genotyping the cohort using Giggles software to determine variant carriers (n = 967 samples), noting that multiallelic sites showed higher Mendelian inconsistency (15.1%).

    Study findings

    The present study resulted in the production of a richly annotated, publicly available catalog of more than 100,000 sequence-resolved SVs (biallelic), alongside 369,685 multiallelic variable number tandem repeats (VNTRs) genotyped using the Vamos tool. Identified SVs included inversions, deletions, duplications, and insertions, totalling a greater than tenfold increase in the number of fully resolved insertion sites, filling a critical gap in human genomic knowledge.

    Mendelian consistency experiments leveraging family trios (two parents and a child) within the cohort demonstrated the study’s high accuracy and extremely low error rate (deletions and insertions at just 3.87% and 4.44%, respectively) for biallelic SVs. Notably, most of the novel SVs identified in this study were found to be extremely rare, with 59.3% having a minor allele frequency (MAF) of less than 1%. Individuals of African descent demonstrated the highest degree of SV diversity.

    Finally, the study provided novel insights into the biological mechanisms that create SVs, detailing how mobile DNA elements, such as L1 and SVA retrotransposons, drive genetic innovation by promoting SV formation and translocation through locus-specific processes, including promoter hijacking (e.g., the 8q21.11 L1 source element).

    Conclusions

    The present study represents a commendable leap forward in our knowledge and understanding of human genomics. The application of long-read sequencing successfully allowed for the discovery and annotation of more SVs (especially insertions), and the diversity of the sample cohort (26 distinct ancestries across several continents) validates the generalizability and global application of study findings.

    Furthermore, the resultant comprehensive and accurate SV atlas, being open access, opens the doors to a new era of genetic medicine, allowing for the identification and early treatment of genetic conditions that we hitherto didn’t even know existed. Notably, when applied to rare-disease genomes, the resource filtered 55% of candidate SVs while retaining 94% (35/37) of validated causal variants. This open-access resource will be invaluable for the scientific community, enabling a deeper understanding of human evolution, population genetics, and the functional consequences of genetic variation.

    Journal reference:

    • Schloissnig, S., Pani, S., Ebler, J., Hain, C., Tsapalou, V., Söylev, A., Hüther, P., Ashraf, H., Prodanov, T., Asparuhova, M., Magalhães, H., Höps, W., Sotelo-Fonseca, J. E., Fitzgerald, T., Santana-Garcia, W., Moreira-Pinhal, R., Hunt, S., Pérez-Llanos, F. J., Wollenweber, T. E., … Korbel, J. O. (2025). Structural variation in 1,019 diverse humans based on long-read sequencing. Nature. DOI – 10.1038/s41586-025-09290-7, https://www.nature.com/articles/s41586-025-09290-7

    Continue Reading

  • Gut bacteria send direct signals to the brain to stop you from overeating

    Gut bacteria send direct signals to the brain to stop you from overeating

    A new study reveals how your gut bacteria can directly tell your brain when to stop eating, opening fresh avenues for appetite and metabolic research.

    Study: A gut sense for a microbial pattern regulates feeding. Image Credit: Pormezz / Shutterstock

    A recent study published in the journal Nature found that feeding behavior is regulated by a previously unknown gut-brain sensory modality for microbial patterns.

    All organisms interpret the world through their senses. A growing body of evidence has established a neural basis of gut sense, which constantly assesses gut luminal stimuli. Epithelial neuropod cells in the small intestine detect nutrients and relay the information via the vagus nerve to influence real-time appetitive choices.

    Evidence suggests that gut microbes can modulate feeding behavior through immune signals, vagal pathways, and neuromodulators. Nonetheless, a direct neural circuit sensing microbial signals to regulate feeding is unknown. Vagal neurons form neuroepithelial circuits with neuropod cells labelled by the satiety-inducing protein, peptide YY (PYY). These PYY-neuropod cells specifically express TLR5, unlike serotonin-producing enteroendocrine cells.

    The neuropod cells and other colonic cells are constantly exposed to microbes that can be recognized by microbial molecular patterns, including flagellin. Flagellin, the structural component of flagella, is conserved across bacterial phyla and activates the Toll-like receptor 5 (TLR5), a pattern-recognition receptor (PRR). Notably, flagellin levels physiologically increase in stool upon feeding.

    The study and findings

    In the present study, researchers determined that a gut-brain sensory modality for microbial patterns termed the “neurobiotic sense” regulates feeding behavior. First, the intestinal epithelial cell transcriptomes were sequenced using reporter mice that expressed the green fluorescent protein (GFP) under the promoters of PYY and cholecystokinin (CCK). The colon and distal ileum are enriched with PYY-labeled cells, while CCK prevails in the proximal intestinal cells.

    The transcriptomes of sensory epithelial cells were compared to those of neighboring epithelial cells to identify microbial signal receptors. The team found significant enrichment of genes encoding microbial byproduct receptors, including G protein-coupled receptor 119 (Gpr119), G protein-coupled bile acid receptor 1 (Gpbar1), free fatty acid receptor 1 (Ffar1), and Ffar2. However, only PYY-GFP cells were significantly enriched for PRRs, with Tlr5 being the most prominent.

    In situ hybridization was performed to verify the expression of Tlr5 in PYY-labeled cells. Colocalization increased from 24% in the ileum to 57% in the distal colon, where PYY-labeled cells have the highest density. Next, Tlr5 was knocked out from PYY-labeled cells in mice. Tlr5 ablation in PYY-labeled cells resulted in mice eating more (increased meal size in both sexes and longer meal duration in females) and gaining more weight than controls, independent of canonical immune signaling (MyD88), metabolic dysfunction, or inflammation.

    Further, the researchers found that the relative levels of flagellin (a TLR5 ligand) in the stool were higher in fed mice than in fasted mice, and this response was unaltered in Tlr5-ablated mice. This indicated that colonic flagellin levels were independent of Tlr5 expression in PYY-labeled cells and that feeding correlated with higher colonic flagellin levels. In addition, the team found that PYY-labeled cells utilize TLR5 to sense flagellin, but not other TLR ligands, such as Poly(I:C), and transduce this signal by releasing PYY. Vagal neurons themselves lack TLR5 and show no direct response to flagellin.

    PYY-labeled cells were found to be significantly enriched for genes involved in synaptic formation, signaling, and neurotransmission. Besides, one-fifth of PYY-labeled cells contact peripheral neurons in the colon and ileum. As such, the researchers sought to confirm whether these cells are connected to the vagus nerve using luminal optogenetics and whole-nerve electrophysiology recordings of the cervical vagus nerve, with intralipid serving as a positive control.

    This showed that PYY-labeled cells directly activate the vagus nerve, establishing a direct signaling circuit between the colon and hindbrain. Further experiments indicated that PYY-labeled neuropod cells sense luminal flagellin using TLR5 and transduce this microbial signal to the vagus nerve. Next, the team investigated whether PYY release from neuropod cells was necessary for vagal activity in response to flagellin stimulus.

    Blocking the neuropeptide Y receptor type 2 (Y2R), the PYY receptor on colonic vagal neurons, ablated cervical activity in response to flagellin. Calcium imaging of vagal nodose neurons revealed that 60.6% of neurons exclusively responded to flagellin, while 27.7% responded to both flagellin and nutrients, indicating the existence of a unique neuroepithelial circuit for flagellin sensing, alongside potential nutrient integration pathways. This discovery prompted investigations into how flagellin influences feeding behavior in real time.

    To this end, mice were fasted overnight to induce hunger and were administered flagellin or phosphate-buffered saline (PBS) by enema. Flagellin enema resulted in a significant decrease in food intake within 20 minutes in littermate controls but had no effect on mice with TLR5 ablation in PYY-labeled cells.

    Pharmacological inhibition of Y2R or TLR5 prevented flagellin-induced decrease in food consumption, suggesting that flagellin reversibly and rapidly suppresses food intake. Finally, the flagellin enema also reduced food intake in germ-free mice, suggesting that flagellin sensing is sufficient to suppress food intake, regardless of microbial signals.

    Conclusions

    In sum, PYY-labeled colonic neuropod cells use TLR5 to detect flagellin and rapidly signal to the brain via the vagus nerve to regulate feeding behavior through dedicated NPY2R receptors. This gut-brain neural circuit forms a neurobiotic sense, enabling the host to adjust behavior by monitoring gut microbial patterns. Notably, the study utilized Salmonella typhimurium flagellin, warranting investigations into other molecular variants of flagellin, as bacteria can be commensal or pathogenic depending on the specific variant expressed.

    Continue Reading