Science Week — #8

What if you could take a census of the invisible—dark energy, dark matter, and distant worlds—without ever leaving Earth? This week, a new space telescope set off to try exactly that. Welcome to episode 8 of our weekly show, the science of the week for September 1. I’m Grace, and I’ve been thinking a lot about how we actually learn about things we can’t see.

And I’m Daniel. It’s a good week to ask that question. We’ve got a major space observatory on its way to a quiet point in space, a genetic study hinting at two previously unknown human ancestor lineages, a computational map of how different tissues age, and some careful reality checks—from Mars images to lab safety.

Also a reminder that “wow” and “wait a second” can coexist. Let’s start with our deep past and how it echoes in our genes.

Scientists at the University of California, Berkeley report they’ve identified DNA from two previously unknown human ancestor lineages in the genomes of modern people. They used a newly developed genomic analysis technique, and one of those lineages apparently interbred with ancestors in Africa more than 50,000 years ago. That’s a mind-bender: traces of branches on our family tree that we didn’t even know were there.

It is intriguing, but let’s keep the footing steady. What we have here is an inference from patterns in modern genomes using a new analysis method. The claim is that signals in the data fit best if you allow contributions from two ancestral lineages that aren’t the usual suspects we already talk about. The detail provided is that one interbreeding event occurred more than 50,000 years ago in Africa.

So we’re not saying “a new fossil was found,” we’re saying “the math suggests missing branches.”

Exactly. No bones on a table—just statistical fingerprints. That’s valuable, but preliminary. We don’t have sample sizes or replication details here, and it’s one research group presenting a new technique. Confirmation would look like independent teams reproducing the patterns with different datasets and methods, plus, if we’re lucky, future archaeological or fossil evidence aligning with the genetic story.

Still, it’s a powerful idea: our genomes as archives, with footnotes written by ancestors we haven’t formally met yet.

Speaking of time written into biology, a team applied a computational pipeline called PathStAR to map how different human tissues structurally age. They analyzed over 25,000 post-mortem histology slides from the GTEx project, spanning 40 human tissues. The result is a set of distinct, nonlinear aging patterns and a picture of coordinated deterioration linked to inflammation, metabolic decline, and hormonal changes.

The scale matters here—more than 25,000 slides across 40 tissues is a substantial dataset. Histology slides are thin tissue sections you can stain and inspect; the “structural aging trajectories” are essentially patterns the algorithm extracts as tissue architecture shifts with age. The emphasis on nonlinear patterns means it’s not just a straight line from young to old; some tissues may change slowly and then accelerate, or vice versa.

The coordinated aspect is interesting too—if inflammation ramps up in one system, others might echo that drift, which could help explain why aging feels like a symphony, not a solo.

A symphony with some dissonance. Important caveat: this is observational and computational. It links tissue structure and known pathways like inflammation and metabolism, but it doesn’t prove that, say, inflammation causes a given structural change. Also, the data are post-mortem slides; that’s a powerful resource, but it lacks the real-time dynamics you’d get from longitudinal imaging.

Still, mapping where and when the cracks appear can guide where we look for causal levers later—and what to measure in trials that try to slow or soften those trajectories.

On to brain health. Researchers at the University of South Carolina used a polymer gel to deliver an antibody that degrades a protein called PTBP1 in astrocytes. Those are support cells in the brain. With PTBP1 reduced, the astrocytes adopted neuronal characteristics in vitro, and the approach led to improved memory-related behavior in mouse models of Alzheimer’s disease. That’s a hopeful signal.

It is a neat mechanistic story: alter PTBP1 in astrocytes and they take on neuron-like features—that’s the “in vitro” part, meaning in the dish. Then, in mice that model aspects of Alzheimer’s disease, delivering the antibody via a gel correlated with better performance on memory-related tasks. But let’s pin the qualifiers in place: mouse models are not humans, and neuron-like does not equal fully functional neurons integrated into a living brain’s circuitry.

So it’s a demonstration of possibility, not a therapy.

Correct. We don’t have details here on dosing, duration of effect, off-target impacts, or safety. Also, “improved memory-related behavior” in mice can be sensitive to experimental conditions. Replication and deeper characterization would be the next steps—plus, if this ever approaches clinical testing, there’s a long road to confirm that converting or reprogramming cells is safe and actually beneficial in human disease.

But as a proof-of-concept—using a delivery gel and an antibody to nudge support cells toward a new role—it’s an interesting avenue to keep an eye on.

A quick lab safety note with a dose of relief. MIT said on August 26 that a graduate student who claimed to have synthesized dimethyl mercury showed no signs of mercury exposure in an initial blood test. The building stayed closed for decontamination, and city emergency responders and industrial-hygiene experts were brought in to assess the risk and ensure safety.

The key word is “initial.” It’s good that the early test showed no exposure, but the responsible thing is to complete decontamination and follow through on all testing and clearance steps. The response—closing the building and bringing in experts—tracks with best practice when dealing with a highly toxic compound.

It’s a reminder that careful protocols and sober communication matter. Science works best when it’s curious and cautious at the same time.

Planet and sky watchers—this is a follow-up. We covered the upcoming partial lunar eclipse earlier, and The Weather Channel’s latest note said the eclipse set for the night of August 27, 2026 would be visible across North and South America, but with mixed cloud forecasts. Clear skies under a heat dome in parts of the Southwest; scattered showers and thunderstorms elsewhere.

Forecasts like that are about probabilities, not guarantees. For an event like a lunar eclipse—safe to watch with the naked eye—the weather is the main variable. The update adds regional texture: good odds in some dry, hot zones; more uncertainty in stormier regions.

If you planned a watch party, the sky might have had the final say. The Moon keeps its appointments; we humans negotiate with the clouds.

All right, to the headline-grabber. NASA launched the Nancy Grace Roman Space Telescope on August 30 aboard a SpaceX Falcon Heavy from Kennedy Space Center. It’s a roughly five-year mission that will set up shop at the Sun–Earth L2 point—quiet, thermally stable, and great for wide-field observations. Roman is built to map dark energy and dark matter, and to census exoplanet populations. That’s a bold to-do list for a single observatory.

The mission cost is cited as about four billion dollars, and the launch followed weeks of preparation, including final inspections before sending a sensitive instrument into space. The science brief—dark energy, dark matter, and exoplanets—translates into deep, wide surveys that can tighten constraints on how the universe expands, how structure forms, and how common different types of planets are. We should expect formal results only after commissioning and calibration; it’s a marathon, not a sprint.

And in the corridor between science and politics, the week also included a salute: on August 31, President Donald Trump called a NASA press briefing to congratulate the agency on the Roman launch. It’s notable, given that the telescope’s funding had previously faced cancellation efforts.

We also saw a separate ceremony on August 28 where President Trump presented the Congressional Space Medal of Honor to the four Artemis II astronauts, recognizing their historic lunar flyby. The report notes it’s the first time the nation’s highest space honor has been awarded for a lunar mission. Those are ceremonial events, not research results, but they reflect how human spaceflight milestones are recognized.

Back on Mars—sort of—NASA addressed a viral image from the Curiosity rover that looked like a tripod-shaped object on the surface. The agency says it’s not a Martian tchotchke; it’s a particle artifact. A high-energy particle hit the rover’s navigation camera and produced the feature. NASA had previously explained a dark silhouette captured by Curiosity was also likely the result of a high-energy particle strike.

That’s the sober lesson: cameras in deep space and on other planets get peppered by energetic particles—cosmic rays and solar particles—that can momentarily light up sensors. The result can look like streaks, specks, or odd silhouettes. It’s not evidence of structures on Mars. If it were physical, you’d expect it to show up in multiple frames or instruments; here, it’s a single-frame artifact.

The universe throws a lot of noise at our cameras. Good thing we have people trained to tell pixels from pebbles.

More space operations news: NASA announced on August 31 that it had reactivated two of the Swift space telescope’s three science instruments after canceling an orbit-boost rescue mission. That cancellation followed an earlier abort on August 26, when the Link satellite built by Katalyst Space Technologies ended its attempt.

So, two of three instruments are back online. That’s partial recovery, not full restoration. We don’t have detail here on which instruments or what capabilities remain limited, but “two of three” is the operative fact. The canceled orbit-boost would have been an operations maneuver; with it off the table, getting instruments working again buys time and science, even if at reduced capacity.

Meanwhile, from ground to sky on a grand scale: the European Southern Observatory says its Extremely Large Telescope is expected to achieve first light by 2030 on Cerro Armazones in Chile’s Atacama Desert. The dome is slated for completion in 2027. A 39-meter mirror is—well—vast. The promise is unprecedented resolution from the ground.

The timeline is an expectation, not an achievement yet. But the milestones—dome completion by 2027, then commissioning toward first light by 2030—give a realistic horizon for when we might start seeing data. Location matters: the Atacama’s dry, stable air is ideal. “Unprecedented resolution” is a fair descriptor for that aperture plus modern correction systems, but we’ll still need to see how performance lines up once the telescope is on sky.

Between Roman at L2 and the ELT in the Atacama, it feels like we’re building complementary eyes: a space-based wide-field mapper and a giant ground-based zoom lens.

Physics and materials, or maybe this week, instruments as the message. We don’t have a fresh particle-physics headline here, but the engineering threads are strong: a 39-meter segmented mirror coming together in the desert, a space telescope tuned for faint signals, and detectors resilient enough to take a beating from high-energy particles—and flag the fakes.

It’s a good reminder that advances in materials, optics, and detectors are the quiet partners in headline science. This week’s facts are about timelines and capabilities rather than new measurements, and that’s fine. Building the tools is a scientific achievement in its own right.

So what did we learn about how the world works this week? That our genomes carry whispers from ancestors we didn’t know we had. That aging isn’t a straight line but a set of curves that bend differently across tissues—curves we can now start to map at scale. That support cells in the brain can be nudged toward neuron-like roles in mice, a hint of possibility that needs time and proof. And that careful process—weather forecasts, instrument reactivations, clear-eyed image analysis—keeps wonder tethered to truth.

And in space science, the arc is long. Roman’s five-year mission will feed cosmology and exoplanet statistics for years. The ELT’s 2030 first-light target shows how multi-year engineering becomes data, then discovery. In the meantime, we celebrate milestones, and we keep separating signal from artifact—literally.

If you want calm, checked science news the rest of the week, visit factslane.com or find us in the App Store and Google Play. And if you like this show, you might also enjoy our weekly companions—Good Week and Tech Week.

Thanks for spending time with us. Stay curious, take care of each other, and we’ll talk again next week.

See you next week.