Decoding Extraterrestrial Organics: GC×GC-HR-TOF-MS and Machine Learning Differentiate Abiotic and Biotic Signatures

A joint study between the Georgia Institute of Technology (Atlanta, Georgia) and the Goddard Space Flight Center at NASA (Greenbelt, Maryland) analyzed a suite of meteoritic and terrestrial samples and systematically compared the molecular distribution of their organic contents using 2D gas chromatography coupled to high-resolution time-of-flight mass spectrometry (⁠⁠GC x GC-HR-TOF-MS). Hypothesizing that meteorite samples can be distinguished from terrestrial rocks by differences in the distributions of their organic compounds, this research led to a paper published in PNAS Nexus (1).

Future sample return missions—such as NASA’s Mars Sample Return (MSR) and JAXA’s Martian Moons eXploration (MMX)—aim to bring samples from the Martian system back to Earth, including material from environments that may once have harbored life. Critical to answering questions regarding the origins of life, these missions require the development of impartial tactics for the assessment of where the line dividing samples containing abiotically synthesized organic compounds from those containing biotic remnants of past or present life should be established (2-4). Included in these samples will be carbonaceous chondrites, which are a primitive and organically rich class of meteorites that represent the oldest solid materials in the Solar System obtainable for scientific analysis. The soluble organic compounds contained in these meteorites provide evidence of abiotic chemical reactions which existed before the beginning of life as we know it (5-7).

“Determining whether organic molecules in planetary samples originate from biological or nonbiological processes,” the authors wrote in their article, “is central to the search for life beyond Earth. Yet, distinguishing these origins is challenging due to overlapping chemical signatures and limited access to pristine extraterrestrial materials.” (1)

To deconvolute the resulting large dataset which resulted from their analysis, the team developed LifeTracer, a computational framework for processing and downstream machine learning (ML) analysis of mass spectrometry data. Leveraging analyte metrics from ⁠GC x GC-HR-TOF-MS, LifeTracer identified predictive molecular features that distinguish abiotic from biotic origins and enabled a robust classification of meteorites from terrestrial samples based on the composition of their nonpolar soluble organics (1).

“In contrast to traditional biomarker-based approaches,” the authors wrote, “LifeTracer analyzes untargeted chemical signatures to infer molecular origin with high accuracy. It enables scalable, unbiased biosignature detection and offers a powerful tool for interpreting complex organic mixtures returned by current and future planetary missions” (1).

LifeTracer identified an unbiased set of discriminative features across sample types, which enabled classification based on their organic distributions, without relying on the presence or absence of known biosignatures. However, scientists cannot rely on single organic molecules to determine whether extraterrestrial organics come from life or non-life, as many of these compounds appear in both. Therefore, overall molecular patterns are compared, which can act as biosignatures. As Martian samples already contain complex mixtures from multiple sources, advanced multivariate tools like LifeTracer—supported by machine learning and expert analysis—will be crucial for assessing whether any patterns resemble known biological or abiotic signatures, especially as Mars sample return missions progress (1).

Read More on Similar Topics:

Laser Thermal Desorption and GC×GC for Harsh Environment and Planetary Mass Spectrometry

Read More on Similar Topics published by our sister publication, Spectroscopy:

Celebrating National Space Day with Spectroscopy

References

1. Saeedi, D.; Buckner, D.; Walton, T. A. et al. Discriminating Abiotic and Biotic Organics in Meteorite and Terrestrial Samples Using Machine Learning on Mass Spectrometry Data. PNAS Nexus 2025, 4 (11), pgaf334. DOI: 10.1093/pnasnexus/pgaf334
2. Grady, M. M. Exploring Mars with Returned Samples Space Sci. Rev. 2020, 216, 51. DOI: 10.1007/s11214-020-00676-9
3. Saeedi, D.; Buckner, D.; Aponte, J. C. et al. Astroagents: A Multi-Agent AI for Hypothesis Generation from Mass Spectrometry Data [preprint]. arXiv 2025, 2503.23170. DOI: 10.48550/arXiv.2503.23170
4. Usui, T. et al. The Importance of Phobos Sample Return for Understanding the Mars-Moon System. Space Sci. Rev. 2020, 216, 49. DOI: 10.1007/s11214-020-00668-9 - citeas
5. Aponte, J. C.; McLain, H. L.; Saeedi, D. et al. Challenges and Opportunities in Using Amino Acids to Decode Carbonaceous Chondrite and Asteroid Parent Body Processes. Astrobiology 2025, 25 (6), 437-449. DOI: 10.1089/ast.2025.0017
6. Glavin, D. P. et al. The Origin and Evolution of Organic Matter in Carbonaceous Chondrites and Links to Their Parent Bodies. In: Primitive Meteorites and Asteroids: Physical, Chemical, and Spectroscopic Observations Paving the Way to Exploration. Elsevier,2018. p. 205–271.
7. Pizzarello, S.; Shock, E. Carbonaceous Chondrite Meteorites: The Chronicle of a Potential Evolutionary Path between Stars and Life. Orig. Life. Evol. Biosph. 2017,47 (3), 249-260. DOI: 10.1007/s11084-016-9530-1