Scientists have detected chemical evidence of life in 3.3-billion-year-old rocks using advanced chemistry and AI.
The findings push back molecular signs of oxygen-producing photosynthesis by nearly a billion years.
Machine-learning models distinguished biological from non-biological materials with over 90% accuracy.
The method could help resolve long-standing debates about early life and guide the search for extraterrestrial biosignatures.
Scientists have detected chemical traces of life in 3.3-billion-year-old rocks, pushing back the molecular record of biology on Earth by some 800 million years — and offering a powerful new tool in the search for life beyond the planet.
In a study published in the journal Proceedings of the National Academy of Sciences, researchers from the Carnegie Institution for Science and partner universities combined advanced chemical analysis with artificial intelligence to uncover faint biological “whispers” preserved in ancient sediments. The team examined more than 400 samples from modern plants and animals to meteorites and billion-year-old fossils to train a machine-learning model to spot the subtle chemical patterns left behind by living organisms.
Robert Hazen, senior staff scientist at Carnegie and co-lead author, in a statement said: “Ancient life leaves more than fossils; it leaves chemical echoes. Using machine learning, we can now reliably interpret these echoes for the first time.”
The researchers used pyrolysis gas chromatography-mass spectrometry to break samples into molecular fragments. Their artificial intelligence model distinguished biological from non-biological materials with up to 98 per cent accuracy and identified photosynthetic signatures in rocks at least 2.5 billion years old, which is nearly a billion years earlier than previously documented.
The findings could transform scientists’ ability to read the deep-time record of life on Earth. Until now, molecular traces that could be confidently tied to biology had only been found in rocks younger than 1.7 billion years, as heat and pressure tend to break down original biomolecules. The new method shows that even heavily degraded organic fragments still contain patterns unique to life.
Anirudh Prabhu of Carnegie Science, a co-first author, said in a statement: “Even when degradation makes it difficult to spot signs of life, our machine-learning models can still detect the subtle traces left behind by ancient biological processes.”
The study also detected molecular evidence of oxygen-producing photosynthesis in rocks 2.5 billion years old. This is an early signal of the process that ultimately transformed Earth’s atmosphere and paved the way for complex life.
Katie Maloney, from Michigan State University, who contributed rare billion-year-old seaweed fossils, said in a statement: “Pairing chemical analysis and machine learning has revealed biological clues about ancient life that were previously invisible.”
Beyond Earth, the authors say the technique could help analyse rocks from Mars or icy moons such as Europa, offering a new way to search for traces of ancient extraterrestrial life.