Every new paper, in one sentence you can actually read.

Each issue takes 15 brand-new scholarly papers, says in a single sentence what each one found, then explains it. Every paper is screened for retractions and a live DOI before it reaches you.

This screen checks process signals. It does not tell you the findings are true.

Get it free, every two days

Five papers free every issue, about 15 emails a month. No spam. Leave in one click.

Prefer no email? The same issues are in the RSS feed.

In this issue 15 papers · 5 free

01 PLOS Biology A microRNA atlas of CD4+ T cell subsets in experimental autoimmune encephalomyelitis identifies regulators of disease pathogenesis Free
02 The Astrophysical Journal An Approximate Bayesian Deep Learning Approach for Uncertainty-aware Differential Emission Measure Estimates in the Solar Corona from the SDO Free
03 Nature Communications Pan-Ebolavirus nanoparticle vaccine provides protection in rodents from lethal infection by Zaire and Sudan viruses Free
04 Nature Communications Automatic value learning results in counterproductive human behavior Free
05 Geophysical Research Letters Attention‐Based Stochastic Simulation of Climate‐Informed Spatiotemporal Flood Risk for Multisite Insurance Portfolios Free
06 Nature Physics Time-resolved imaging of antiferromagnetic skyrmion interactions Subscribers
07 Nature Human Behaviour Hierarchical resource rationality explains human reading behaviour Subscribers
08 eLife A tool to pulse-label yeast nuclear pore complexes in imaging and biochemical experiments Subscribers
09 Science Advances A clinical and experimental investigation of liraglutide effects on the brain-kidney axis Subscribers
10 Science Advances Language-like statistical structure arises in learned signaling: evidence from birdsong Subscribers
11 Science A central role for dietary sugars in human evolution Subscribers
12 Nature Communications Synergistic effect of paired Cu(II) open metal sites for enhanced high-temperature hydrogen isotope separation Subscribers
13 Proceedings of the National Academy of Sciences Lipid droplets confer SA-dependent antiviral defense without growth penalty Subscribers
14 Proceedings of the National Academy of Sciences An antisense method for efficient exon skipping and its application to Duchenne muscular dystrophy Subscribers
15 PNAS Nexus Measuring narrative polarization in online discourse Subscribers

The papers Read free

PLOS Biology·10 August 2026

Carolina Cunha, Daniel Inácio et al.

In mice, two gene switches normally curb harmful immune cells but fail inside inflamed brain tissue.

Two tiny molecules that control genes help hold back harmful immune cells, and this control breaks down during brain inflammation. In mice with a multiple sclerosis like disease, scientists mapped which small gene switches turn on in different types of immune helper cells. They found 110 of these switches differ between inflammatory cells and calming regulatory cells. Two of them, when studied closely, changed how fast inflammatory cells multiplied and how they matured, affecting how severe the disease became.

The number 110 microRNAs

The number of small gene switches found to differ between inflammatory and calming immune cells in mice.

Mouse model only, no human data yet.

Integrity screen: passed (3 checks) Checked 17 August 2026. Carolina Cunha, Daniel Inácio et al., A microRNA atlas of CD4+ T cell subsets in experimental autoimmune encephalomyelitis identifies regulators of disease pathogenesis Read the source

The Astrophysical Journal·10 August 2026

N. Balodhi, R. J. Morton

In sun images, a new AI model estimates coronal temperature fast and reports its own confidence.

A computer model can quickly guess how hot the gas is in the sun's outer atmosphere. It uses pictures from a NASA space telescope. The model also tells scientists how sure it is about each guess, spot by spot. It matched older, slower methods but worked much faster.

What it could change A fast AI tool estimates sun temperature maps and flags its own uncertainty per pixel.

Shows speed and built-in confidence scores, not a new physical discovery about the sun.

Model only, tested against older calculation methods, not new sun physics.

Integrity screen: passed (3 checks) Checked 11 August 2026. N. Balodhi, R. J. Morton, An Approximate Bayesian Deep Learning Approach for Uncertainty-aware Differential Emission Measure Estimates in the Solar Corona from the SDO Read the source

Nature Communications·8 August 2026

Connor Weidle, Natalie Brunette et al.

In rodents, a tiny protein cage vaccine protected against two deadly Ebola virus types.

A lab-made protein shell shaped like two different Ebola viruses at once protected animals from deadly infection. Mice given a shot matched to one virus survived best. Guinea pigs given a mixed shot survived a related virus and made antibodies against both viruses. The mixed vaccine worked worse than a matched single-virus shot in mice, so results differed between species.

What it could change A single vaccine shell shaped like two Ebola virus types protected rodents from both.

Shown only in mice and guinea pigs, not yet tested in humans.

Mice and guinea pigs only, no human data yet.

Integrity screen: passed (3 checks) Checked 11 August 2026. Connor Weidle, Natalie Brunette et al., Pan-Ebolavirus nanoparticle vaccine provides protection in rodents from lethal infection by Zaire and Sudan viruses Read the source

Nature Communications·10 August 2026

Ido Ben-Artzi, Maayan Pereg et al.

People automatically assign value to useless details, hurting their choices even when warned.

People keep learning value from things that do not matter, even when they know better. In several lab experiments with more than 1,200 people total, people kept treating random features, like hand movements or visual details, as if they predicted rewards. This happened even after they were told clearly these features meant nothing. People with weaker working memory, the mental space used to hold information in mind, showed more of this unhelpful learning.

The number 504, 158, 195, and 319 people across four experiments

The number of people tested in each experiment showing this automatic, unhelpful learning pattern.

Lab tasks only, unclear if this affects everyday choices like shopping.

Integrity screen: passed (3 checks) Checked 11 August 2026. Ido Ben-Artzi, Maayan Pereg et al., Automatic value learning results in counterproductive human behavior Read the source

Geophysical Research Letters·7 August 2026

Adam Nayak, Pierre Gentine et al.

A new tool predicts flood clusters across many river sites for the next 1 to 10 years.

A new computer model creates realistic flood forecasts for many river sites at once. It focuses on the 1 to 10 year time span that matters most for flood insurance contracts. The model learns from past weather patterns that match today's climate, then generates likely sequences of flood timing, size, and length. Tested on more than 100 sites in the Mississippi River Basin, it found that floods often hit many places at the same time, linked to known large climate patterns.

The number 100+ sites

The number of Mississippi River Basin locations used to test the flood simulation model.

Tested only in one US river basin, not globally validated yet.

Integrity screen: passed (3 checks) Checked 11 August 2026. Adam Nayak, Pierre Gentine et al., Attention‐Based Stochastic Simulation of Climate‐Informed Spatiotemporal Flood Risk for Multisite Insurance Portfolios Read the source

10 more papers are in the full issue.

The same screening, the full reading, every two days. 2 euros a month, or 20 a year, and you can leave in one click.

Nature Physics·10 August 2026

Mona Bhukta, Takaaki Dohi et al.