Oldest Chemical Evidence of Life on Earth Found Inside 3.3 Billion Year Old Rocks

Using AI and advanced chemistry, scientists uncovered 3.3-billion-year-old chemical traces that mark the earliest evidence of life on Earth.

Scientists at the Carnegie Institution for Science used a powerful combination of cutting-edge chemistry and artificial intelligence to peer through the geological fog of time in search of the very first signs of life. They found the earliest and most confident chemical evidence of life on Earth from 3.33-billion-year-old sediments inside ancient rocks from South Africa. This evidence includes fragmentary traces of carbon. 

The new findings essentially double the window of time that organic molecules preserved in rocks can offer useful information about early life. Before this study, no such molecular traces had been reliably found in rocks older than about 1.7 billion years.  

The researchers also identified a stunningly early sign of photosynthesis inside rocks from South Africa and Canada, roughly 2.5 billion years old. It pushed back the chemical record of this world-changing process by over 800 million years.  

The Challenge of Finding Ancient Life

Tracing back life’s earliest steps requires thinking outside the box. Primitive, microscopic organisms don’t fossilize like dinosaur bones and ancient Earth looked radically different than today. About three billion years ago, Earth as dominated by intense volcanic activity, making it a fiery and hellish landscape with molten magma, lava rivers, and thick plumes of volcanic gases like carbon dioxide and methane. 

The first creatures have long been buried, crushed, heated, and fractured within Earth’s crust, nearly obliterating the vital clues to the origins of life — but not entirely. 

While the vast majority of carbon-rich sediments have been altered to the point where even the hardiest organic molecules are broken down into small, generic fragments, modern tools may still piece together a convincing picture from these ancient carbon fragments. 

Carnegie researchers Robert Hazen, Michael Wong, and Anirudh Prabhu, had a different idea from other studies. Their hypothesis is that life’s molecules are rigorously selected for their biological functions, meaning they don’t appear in random distributions like those found in meteorites. Even when the original, intact biomolecules are gone, the fragments they leave behind might still preserve a chemical pattern unique to life.

Finding Life’s Pattern

“Think of it like showing thousands of jigsaw puzzle pieces to a computer and asking whether the original scene was a flower or a meteorite,” Hazen explained in a press release. He added, “Rather than focus on individual molecules, we looked for chemical patterns, and those patterns could be true elsewhere in the universe.”

RelatedPosts

Graphene transistors made using DNA assembly
New capture technology scrubs atmospheric CO2 on the cheap
World’s Tropical Forests are Getting Younger And That’s Actually a Major Problem for the Climate
NASA scientists create basic building blocks of life in a ‘primordial ocean’

To test this, the scientists analyzed over 400 samples. This massive sample set included modern plants, animals, fungi, fossil materials like coal, carbon-rich meteorites, synthetic organic materials, and ancient sediments. Using sophisticated spectrometry, they released the trapped chemical fragments from each sample. They then fed this data to a specific machine learning model called random forest. They then trained the model to recognize the chemical “fingerprints” left behind by biological processes.

This powerful combination proved incredibly accurate. The model could distinguish between materials of biological origin and non-living origins (like meteoritic or synthetic carbon) with over 90 percent accuracy. In fact, on known samples, it achieved up to 98 percent accuracy. 

“Our results show that ancient life leaves behind more than fossils; it leaves chemical ‘echoes,'” Hazen said. “Using machine learning, we can now reliably interpret these echoes for the first time.” 

A New Timeline for Photosynthesis

The machine learning tool didn’t stop at finding the earliest chemical signs of life. It also provided molecular evidence of one of life’s most transformative evolutionary achievements: oxygen-producing photosynthesis. This is the process used by plants, algae, and many microorganisms to harness sunlight for energy.

The model was able to detect the unique chemical patterns associated with photosynthetic life with 93 percent accuracy. Using this capability, the team identified these signatures in rocks as old as 2.52 billion years from South Africa’s Gamohaan Formation, and also in 2.3-billion-year-old rocks from Canada. This finding dramatically extends the known chemical record of photosynthesis preserved in carbon molecules by more than 800 million years.

“Understanding when photosynthesis emerged helps explain how Earth’s atmosphere became oxygen-rich, a key milestone that allowed complex life, including humans, to evolve,” noted astrobiologist Wong. 

The fact that the model could successfully extract this kind of information from “messy, degraded chemical data” opens up exciting new possibilities. If AI can reliably detect these molecular “ghosts” that survived billions of years of Earth’s turmoil, the same technique could be adapt to air in the search for extraterrestrial life. For instance, the method could be applied to Martian rocks or samples from Jupiter’s icy moon Europa. To this aim, the researchers plan to refine their models, perhaps even testing them on anoxygenic photosynthetic bacteria as possible analogs for organisms beyond Earth.

The findings appeared in the Proceedings of the National Academy of Sciences.