A new artificial intelligence system will tackle one of science’s most difficult questions: “Did this chemistry originate from life?”
The technology, built by researchers at Georgia Tech and NASA’s Goddard Space Flight Center, studies meteorites and Earth’s soil for patterns that may indicate biology.
The system, called LifeTracer, is a machine learning tool that compares complex mixtures of organic molecules.
In tests, LifeTracer accurately distinguished between lifeless space rocks and life-bearing Earth samples about 87% of the time, using only chemical data.
Why alien chemistry confuses us
To build its training set, a team led by Amirali Aghazadeh, an assistant professor of electrical and computer engineering at Georgia Tech, measured soluble organic matter in eight carbon-rich meteorites and 10 terrestrial soils and shales.
Experts relied on mass spectrometry, a technique that identifies molecules by their mass and fragment patterns.
Both space rocks and Earth’s soil contain rich organic chemicals, and distinguishing life-derived molecules from raw chemicals is not as easy as examining single compounds.
Abiotic organic matter created by nonliving processes in space and on rocks includes amino acids and nucleobases that previous studies had already discovered inside carbon-rich meteorites.
LifeTracer converts peaks into patterns
Rather than tracking single molecules, LifeTracer reads a series of signals that a mass spectrometer records as peaks in a two-dimensional chromatogram.
In this study, the team recorded 9,475 different fragment peaks in the meteorite extract and 9,070 different fragment peaks in the earth sample.
From these peaks, the software constructs thousands of features that encode each compound’s mass and residence time in each column of the gas chromatograph.
Logistic regression is a statistical method that estimates the probability between two alternatives and serves as a central classifier that assigns each sample to either the abiotic or biotic class.
Build chemical fingerprints quickly
Because many of these peaks originate from fragments of the same parent molecule, the researchers clustered them into groups that shared similar retention times.
Each group acts like a chemical fingerprint, marking a family of molecules that tends to appear in space rocks and Earth’s materials.
The algorithm then learns which fingerprints bring the sample closer to the non-biological label and which fingerprints bring the sample closer to the biological label.
Once that training is complete, the model can examine entirely new patterns of peaks and determine which aspects are most similar.
Clues buried in ancient space rocks
Many of the meteorites in the dataset belong to carbonaceous chondrites, older stony meteorites that are rich in carbon and preserve early Solar System material.
Previous studies have shown that such meteorites retain tens of thousands of different organic molecules, far more than could be captured by older analytical methods.
In the LifeTracer analysis, fragment ions from meteorites tended to exit the first column of the chromatograph faster than fragment ions from Earth samples. This pattern means that the abiotic mixture was more volatile overall, so its molecules moved faster through the heated column.
Chemistry representing life
Among the most informative fingerprints were polycyclic aromatic hydrocarbons (PAHs), ring-shaped molecules made by bonding carbon atoms and hydrogen that often occur in high-energy environments.
One simple member of the family, naphthalene, repeatedly appeared in meteorite samples and became the strongest single predictor of abiotic origin in the model.
On the Earth side, a particularly distinctive fingerprint is from a benzene ring with multiple long carbon branches, found primarily in soils such as Utah, Atacama, and Iceland.
Their structure matches the type of long, branched molecules that modern organisms use in their cell membranes, so their presence increased the likelihood that the sample would be classified as living.
A place where patterns suggest life
Scientists refer to chemical clues that could point to life as biosignatures, signs that past or present life has shaped a planet’s chemistry.
Many research teams now argue that the most reliable biosignatures come from whole patterns of organic matter, rather than from single specialized molecules.
“Determining whether organic molecules in planetary samples originate from biological or non-biological processes is central to the search for extraterrestrial life,” said study lead author Daniel Saidi of the Georgia Institute of Technology.
His team designed LifeTracer specifically to handle noisy and incomplete datasets, such as those from planetary exploration missions that are expected to return in the future.
LifeTracer for future space missions
Sample return missions from asteroids and planets bring in very small amounts of material, just a few grams of dust or rock scraped from carefully chosen locations.
These samples contain a mixture of organic matter from a variety of sources, including cosmochemistry, surface weathering, and the biology of what may have once lived there.
Future projects such as the Mars Sample Return Campaign and Japan’s Mars Satellite Mission aim to bring back material from potentially habitable areas.
Tools like LifeTracer can help mission scientists quickly classify these mixtures and indicate which mixtures most closely resemble life-modifying chemistry rather than a purely abiotic baseline.
find new clues about life
Because LifeTracer works on a complete distribution of molecules rather than a short list of known biomarkers, it can capture unknown types of chemicals that appear to be organized like life.
This broad view alone cannot prove that life existed elsewhere, but it can lead to more detailed, slower analyzes that look more closely at specific molecules and isotopes.
If scientists can combine pattern-discovery tools like LifeTracer with agents that suggest and test hypotheses, they may finally have a practical way to search huge datasets for subtle signs of biology.
This combination could allow future missions to treat every returning dust particle as a potential clue to how life begins, both on Earth and far beyond.
The research will be published in a journal PNAS Nexus.
—–
Like what you read? Subscribe to our newsletter for fascinating articles, exclusive content and the latest updates.
Check us out on EarthSnap, the free app from Eric Ralls and Earth.com.
—–
