Too Much Pol II, Ghost Supernovae, and Batteries by the Billion

Sponsor: RefurbMiner ERP

0:03Host 1Support for the Hastrup Review of Science comes from RefurbMiner.

0:07Host 1One intelligent platform for your whole business. Invoicing, banking, and compliance in a single system, with every decision fully traceable. Privacy ready, and hosted in Europe.

0:19Host 2Start your free fourteen day trial today. The link is in the show notes.

Cold open

0:24Host 1Welcome to the Hastrup Review of Science — where artificial intelligence reads real science, so you don't have to pretend you understood the paper.

0:33Host 2Five papers this week. One of them rewrites how we think about getting old.

0:37Host 1This week, a team figured out that aging may be driven by having too much of a good thing at birth. We have also got axions constrained by the ghosts of every supernova that ever happened, and some genuinely clever battery science.

0:51Host 2Plus a centenarian gene that turned out to be four genes, and a superconductor that just broke through a major pressure barrier. Five papers, let's get into it.

Aging: the Pol II overshoot

1:00Host 1Del Carmen-Fabregat and colleagues at the Centre for Genomic Regulation studied RNA Polymerase two, the enzyme that transcribes DNA into messenger RNA. They found something surprising in C. elegans: when the animals enter adulthood, they carry about four times more Pol two protein than they actually need.

1:19Host 2That excess is not random. It is a developmental leftover, and as the surplus Pol two slowly decays toward its thermodynamic equilibrium, mRNA abundance drops across most cell types. That decline in total mRNA directly drives the functional deterioration we call aging.

1:36Host 1The manipulation experiments are where this gets really compelling. When they briefly accelerated Pol two loss using targeted interventions, animals aged faster in a clean dose-dependent relationship.

1:48Host 2And when they decelerated the decline, both healthspan and lifespan extended. The conceptual shift is significant: aging here is not caused by damage accumulating over time, it is driven by a developmental overshoot correcting itself back to equilibrium.

2:04Host 1Is this just a worm story or is there reason to think it matters for mammals?

2:09Host 2They confirmed the same pathological decline in absolute mRNA abundance in mammalian tissues during aging, which suggests the mechanism may be conserved. It is a bioRxiv preprint, not yet peer-reviewed, and the interventional data is entirely in worms. The obvious next question is whether you can slow the Pol two decline pharmacologically in a mammalian system, because if you can, that is a genuine longevity target.

TMEM106B: four haplotypes

2:35Host 1Staying with longevity. The TMEM106B gene has been linked to Alzheimer's risk for years through GWAS hits, but Salazar and colleagues just showed the locus is far more complicated than anyone appreciated.

2:47Host 2They integrated Alzheimer's GWAS summary statistics from roughly nine hundred and seventy-eight thousand individuals with deep haplotype analysis in five thousand eight hundred and seventy-three Dutch genomes. That cohort included Alzheimer's cases, age-matched controls, and critically, cognitively healthy centenarians.

3:07Host 1Instead of the two haplotypes everyone assumed were the whole story, they found four.

3:13Host 2Exactly. Two conditionally independent linkage disequilibrium blocks define four haplotypes, T one through T four. The headline is that T three is strongly enriched in centenarians who maintained their cognition, and long-read sequencing in four hundred and ninety-three individuals revealed that T three carries a nineteen kilobase genomic rearrangement plus distinct methylation patterns.

3:37Host 1So the biology is driven by structural variants, not just the single amino acid swap everyone was studying?

3:44Host 2That is the implication. It is a medRxiv preprint and the cohort is entirely Dutch, so replication in other populations is essential. But it moves this gene from a simple risk versus protective binary to a four-haplotype system where structural variation and epigenetics may be doing the heavy lifting.

Ghost supernovae and axions

4:02Host 1Shifting to physics, where Super-Kamiokande recently reported a two point six sigma hint of the diffuse supernova neutrino background, which is the accumulated neutrino glow of every core-collapse supernova in cosmic history. You could call them the ghosts of stars past.

4:19Host 2You could. You should not.

4:21Host 1Mori and colleagues used that signal to hunt for axions, hypothetical particles that could account for dark matter. The logic is elegant: if axions couple to protons, they would carry energy away from collapsing stellar cores, cooling the protoneutron star faster and reducing the neutrino flux we ultimately detect.

4:40Host 2They ran long-term neutrino-radiation hydrodynamic simulations of core-collapse supernovae and compared predictions to the measured diffuse flux. The result is a one-sigma upper limit on the axion-proton coupling constant of one point three times ten to the minus nine, which directly constrains how strongly axions can interact with ordinary matter.

5:02Host 1How does that compare to existing limits?

5:05Host 2It is comparable to the constraint from supernova nineteen eighty-seven A, but the methodological advantage is significant. The nineteen eighty-seven A bound relies on one observed event with all its astrophysical uncertainties, whereas the diffuse background integrates over the entire cosmic population of supernovae. The caveat is that two point six sigma is suggestive, not a discovery, so the constraint tightens only if that signal firms up with more data.

A word from our sponsor

5:33Host 1A word from this week's sponsor. Ladies, do you dread that time of the month? Then you are in luck.

5:40Host 1The Hastrup Review of Science is brought to you by tampons. Cylindrical, absorbent, dependable. Very reasonably priced, and in stock right now at Med24.

5:50Host 2I keep a box in every bag I own. Never once been let down.

5:54Host 1Tampons hold up under pressure, so you don't have to. That's absorbency you can build a whole week around.

6:01Host 2They also fit in a toolbox. Ask me how I know.

6:05Host 1Our referral link is in the show notes. Tampons. They work. Now, back to the science.

Sodium-ion by the billion

6:11Host 1For anyone thinking about energy storage, whether you are running a homelab on solar or just tracking where battery chemistry is headed, sodium-ion is the single most important alternative to lithium right now. Sodium is literally everywhere, it is cheap, and the main bottleneck has been the anode material.

6:29Host 2Hard carbons are the leading anode candidate, but their performance is governed by a messy internal structure, a disordered tangle of graphitic domains, defects, and nanopores. The relationship between that microstructure and actual electrochemical capacity has been frustratingly unclear.

6:46Host 1McLean and colleagues attacked this head on. They generated over thirteen thousand realistic hard carbon structural models using machine-learned interatomic potentials, with individual models containing up to four thousand three hundred and seventy-eight atoms.

7:01Host 2These are not idealized toy structures. They match experimental measurements of density, porosity, and the ratio of sp two to sp three bonding. They simulated sodium intercalation in representative models and reproduced the characteristic sloping-to-plateau voltage profiles seen in real lab experiments.

7:19Host 1Then they trained a lightweight neural network to predict capacity directly from the carbon structure, skipping the expensive full intercalation simulation.

7:29Host 2The surrogate identified candidates exceeding eight hundred milliamp hours per gram, validated by full calculations, compared to roughly three hundred for current commercial hard carbons. I appreciate that they checked against experimental voltage profiles, because too many screening papers generate pretty candidates but never test them against real electrochemistry.

7:51Host 1For someone running batteries at home, where does this leave the sodium-ion timeline?

7:56Host 2It accelerates rational anode design rather than the trial-and-error the field has relied on. If sodium-ion is going to compete with lithium iron phosphate for stationary storage, this kind of validated framework is exactly what was missing.

A superconducting hydride

8:10Host 1Last one, and hydride superconductors have been making headlines with materials like lanthanum hydride superconducting at minus two hundred and thirty Celsius, but they all need crushing pressures above a hundred and fifty gigapascals. Wu and colleagues just changed that equation, you might say they took the pressure off.

8:29Host 2Moving on to the actual result.

8:32Host 1They synthesized magnesium rhodium hexahydride, which superconducts at just thirty gigapascals with a critical temperature of twenty-four kelvin. That rises to twenty-nine kelvin at fifty-three gigapascals. Roughly five times lower pressure than where the field has been stuck.

8:48Host 2The synthesis uses a clever two-step precursor process. First they build a compound with five hydrogens per formula unit stabilized by covalent bonds at lower pressure. Then they add hydrogen, filling antibonding orbitals and triggering a structural transition from a square pyramid to an octahedron around the rhodium, with superconductivity switching on right at that transition.

9:11Host 1Twenty-four kelvin is still very cold, obviously. Is the pressure reduction the real story here?

9:17Host 2It is. Thirty gigapascals opens far more experimental design space than a hundred and fifty, and the precursor strategy is potentially a reusable template: pre-build part of the hydrogen framework with covalent chemistry at manageable pressures, then let pressure do the final rearrangement.

9:34Host 1A recipe that could work for other compositions too?

9:37Host 2Potentially. One compound does not make a trend, but if it generalizes, it could meaningfully change how the field searches for lower-pressure hydride superconductors.

9:48Host 1That is the week. Aging driven by a Pol two overshoot, centenarian genetics revealing four haplotypes where we thought there were two, axions constrained by the ghosts of every supernova that ever happened, sodium battery anodes getting the machine-learning treatment, and a superconductor at a fifth of the usual pressure.

Sign-off

10:07Host 2This has been the Hastrup Review of Science. See you next week.