Summary: A Cambridge team says JWST detected possible biosignatures on exoplanet K2-18b, but an Oxford re-analysis challenges those findings. The disagreement reveals just how hard it is to pull molecular signals from noisy telescope data, and why no one should be declaring alien life just yet.
NASA's Kepler mission discovered K2-18b in 2015, and today that distant world sits at the center of one of the most heated debates in exoplanet science. A team at Cambridge says they found signs of molecules that, on Earth, only living things produce. Another researcher at Oxford says the data shows nothing at all. And the truth? It is buried in the noise.
Cambridge Team Detects Possible Biosignatures on K2-18b
K2-18b weighs about 8.6 times Earth's mass and stretches roughly 2.7 times our radius. Different sources place it between 110 and 124 light-years away, a reminder that even basic measurements of distant worlds carry real uncertainty. The planet orbits its star in the habitable zone, where temperatures could allow liquid water to exist. Scientists classify it as a sub-Neptune, a type of planet with a thick, hydrogen-rich atmosphere that does not exist in our solar system.
The Cambridge team, led by astrophysicist Nikku Madhusudhan, used JWST's Mid-Infrared Instrument (MIRI) to study K2-18b. They confirmed methane and carbon dioxide in the atmosphere, which JWST had already picked up using other instruments. They also found low ammonia levels. But the headline-grabbing claim was the possible detection of dimethyl sulfide (DMS) and dimethyl disulfide (DMDS). On Earth, only living organisms like algae produce these molecules. The team first spotted hints of DMS in 2023 and published the fuller DMDS findings in Astrophysical Journal Letters.
Why the Doubts Matter
Here is the problem. DMS and DMDS being biological on Earth does not make them biological everywhere. We have exactly one example of a living planet, and drawing universal rules from a sample size of one is risky science.
Jake Taylor at the University of Oxford uploaded a re-analysis to arXiv that has not been peer-reviewed. Taylor ran a basic statistical test on the same JWST data without assuming specific molecules were present. His results returned flat lines rather than the bell-shaped curves that would indicate molecular detections, meaning no molecular signals at all.
Taylor also argued that the Cambridge team should not have compared K2-18b's atmosphere to nitrogen-rich models like Earth's. Madhusudhan pushed back, saying Taylor's models are too simple to capture the real complexity of the data.
The Noise Problem Behind the Debate
What makes this disagreement so important is not just K2-18b. It is the fundamental challenge of pulling tiny molecular signals from JWST data. When you are looking at a planet more than a hundred light-years away, the telescope is not taking a clean photo of the atmosphere. It is measuring minute changes in starlight as the planet passes in front of its star, then teasing apart which wavelengths got absorbed by which molecules.
That process involves layers of modeling assumptions. Change the model, change the assumptions, and you can get wildly different results from the same dataset. One team sees DMS and DMDS. Another sees flat lines. Neither result is necessarily wrong. They are just using different mathematical lenses.
Even the distance to K2-18b shows how messy the underlying data can be. Different sources cite the planet at 110, 120, and 124 light-years away. If we are still refining something as basic as distance, imagine the uncertainty around trace atmospheric molecules.
What Comes Next for K2-18b
Extraordinary claims require extraordinary evidence, and right now the evidence for life on K2-18b is nowhere close to extraordinary. The Hubble Space Telescope made history by detecting water vapor in K2-18b's atmosphere, the first such detection on a habitable-zone exoplanet. JWST has clearly moved the science forward by confirming methane and carbon dioxide. But jumping from "interesting chemistry" to "we found alien life" is a leap the data simply cannot support yet.
Taylor's preprint needs peer review. The Cambridge team's models need independent testing by other groups. As Taylor himself noted, the signal-to-noise ratio is not definitive, and more JWST observations would be needed to get a cleaner result.
K2-18b is teaching us something valuable. The hard part of hunting for alien life was never going to be pointing a telescope. It was always going to be agreeing on what the telescope actually saw. So what do you think it will take for scientists to reach a real consensus on a biosignature detection?
Comments