MK ATLAS· Science Atlas

Today in Science

8 November 2026 (Sun)

We bring a study showing that, of the two conditions for superconductivity, electrons pairing up can happen even at room temperature, and the real obstacle is getting the pairs to move in step; a study finding that dry ground, high pressure and heat egg one another on to make the worst heatwaves 10% hotter; a new Harvard project to build and test an AI for cancer research that forms hypotheses and suggests experiments; and a study that read many health signals from the way people walk.

Physics

Electrons pair up even at room temperature — the real obstacle to superconductivity is keeping in step

Researchers from Utrecht University, Leiden University and Munich measured directly that, in a copper-oxide high-temperature superconductor, electrons already pair up at temperatures far above the point where superconductivity begins.

When electricity flows through a wire, some of its energy is lost as heat because of resistance. A superconductor is a material in which this resistance drops to zero. The trouble is that most superconductors only work when cooled to around minus 260 degrees. In the 1980s, copper oxides were found that become superconducting at much warmer temperatures, but these were still around minus 170 degrees, and why it happens is still poorly understood.

Superconductivity needs two things: two electrons forming a pair, and all those pairs moving together in step. It is like ten couples on a dance floor: it is not enough for each couple to dance, every couple must move to the same beat. In ordinary superconductors that work at low temperatures, both happen at once.

The researchers built their own measuring equipment that can tell whether electricity is flowing one electron at a time or in pairs. Measuring a copper-oxide superconductor with it, they found that the electrons were already paired at temperatures far above the point where superconductivity starts. The researchers think the force that holds the pairs together is strong enough to last even at room temperature, and say the obstacle lies in getting the pairs to keep in step.

Through the eyes of the Atlas

Ohm worked out how well electricity flows by splitting it into the push and the resistance. Superconductivity is the state in which that resistance becomes zero. The Library's piece on the room-temperature superconductor looks at how hospital MRI machines, fusion magnets and quantum computers are all tied to minus 269 degrees. This study narrows down where the key to easing that cold might lie.

A word of cautionNo room-temperature superconductor has been made. The study showed that one of the two conditions can be met even at room temperature; how to get the pairs to keep in step is still unknown.

Source · Science

Earth

Why are the worst heatwaves even hotter? Dry ground, high pressure and heat egg one another on

Researchers at the Potsdam Institute for Climate Impact Research and the International Institute for Applied Systems Analysis showed that once temperatures pass a certain line, dry ground, high pressure and heat egg one another on, making the most extreme heatwaves about 10% hotter.

When high pressure settles over an area, the clouds clear and sunlight heats the ground more. When it is hot, the soil dries out, and dry soil loses its power to cool the ground by evaporating water, so the ground gets hotter still. It has long been known that heat and dry ground egg each other on.

Using machine learning and statistical analysis, the researchers looked at recent mega heatwaves that broke records by a wide margin. They found that from the moment temperatures pass a certain line, the three things, dry ground, high pressure and heat, become tangled together and push one another even higher. This tangle made the most extreme heatwaves about 10% hotter than they would otherwise be. The same thing showed up across the mid-latitudes, in Europe, southern Russia and Mongolia, North America, north-western India and inland Australia.

In Europe, the tangle starts at about 4 degrees above the average summer temperature, roughly 29 degrees. The chance of temperatures in Central Europe passing this line was 6% in 1979 to 1983, and it rose to 23% in 2018 to 2022.

Through the eyes of the Atlas

While calculating the weather, Lorenz found that a tiny difference could grow into completely different weather. In the weather, many factors push and pull one another, and looking at just one of them on its own can miss things. This study points out that one such tangle is not properly included in today's climate models.

A word of cautionThe 10% applies to the most extreme heatwaves, not to all hot weather. The researchers also add a condition: if this reading is right, future heatwave risks may be greater than expected now.

Source · Science Advances

Technology

An AI that forms hypotheses and suggests experiments — could it speed up cancer research?

A team led by Professor Marinka Zitnik at Harvard Medical School received about 7 million dollars for AURORA, a project to build and test an AI co-scientist that helps with cancer research.

Today's AI has become fairly good at finding biological results that are already known. But an AI that forms new hypotheses on its own, checks them by experiment and learns again from the results is still taking its first steps.

The AI the team wants to build gathers and weighs papers and databases, other biological AI tools such as ones that predict protein structures, and results fresh from the laboratory. It then puts forward hypotheses worth testing and suggests experiment plans that fit the conditions the researchers care about, such as time and budget. When the results come in, it reads them and suggests the next direction. The idea is to cut down on dead ends by having scientists look at the most likely candidates first, instead of switching off thousands of genes one by one.

As part of the project, a further 4 million dollars will go to three cancer research teams in Spain, the United Kingdom and the United States to test how much this AI actually helps their research. Scientific judgement and decisions stay with people throughout.

Through the eyes of the Atlas

Turing asked what a machine that only follows rules can and cannot do. Today's question is a little different: how much of a scientist's work can a machine share? This project is an attempt to measure that question in real laboratories.

A word of cautionThis news is not a research result but a project just starting with new funding. Whether this AI really speeds up cancer research is yet to be tested.

Source · Harvard Medical School

Life

How many health signals can be read just from the way someone walks?

Researchers at the Weizmann Institute of Science in Israel used AI to analyse how 3,328 people walk and move, and meaningfully predicted 1,887 of 3,210 health measures.

The way we walk has long been seen as a mirror of health, but it has usually been treated only as a symptom of particular diseases. Because it was measured with just a few numbers chosen by people, such as stride length and walking speed, the fine detail of how each person moves was easy to miss.

The researchers used a single depth camera to film 3,328 participants in the Human Phenotype Project, whose bodies had been examined in great detail, doing five movements. They then had an AI turn the way all the body's joints move together over time into sets of numbers. These numbers predicted age, body mass index and the amount of fat around the organs fairly well, and meaningfully predicted 1,887 of 3,210 measures across 18 body systems, such as grip strength and sleep quality.

Compared with using only basic information such as age, body mass index and height, the predictions improved in all 18 systems for men and in 17 for women. Leg movements carried more about metabolism and frailty, and torso movements more about sleep and lifestyle.

Through the eyes of the Atlas

The Hippocratic school chose observation over prayer and wrote down a patient's fever, sleep and breathing every day. This study, which tries to read the way we walk as a signal from the whole body, widens that observation with a camera and AI.

A word of cautionMeaningful prediction does not mean it can diagnose disease, and more checks are needed before it can be used in real care. Employees of the company that made the filming system took part as co-authors.

Source · Nature Medicine