A robust, agnostic molecular biosignature based on machine learning.
Henderson James CleavesGrethe HystadAnirudh PrabhuMichael L WongGeorge D CodySophia EconomonRobert M HazenPublished in: Proceedings of the National Academy of Sciences of the United States of America (2023)
The search for definitive biosignatures-unambiguous markers of past or present life-is a central goal of paleobiology and astrobiology. We used pyrolysis-gas chromatography coupled to mass spectrometry to analyze chemically disparate samples, including living cells, geologically processed fossil organic material, carbon-rich meteorites, and laboratory-synthesized organic compounds and mixtures. Data from each sample were employed as training and test subsets for machine-learning methods, which resulted in a model that can identify the biogenicity of both contemporary and ancient geologically processed samples with ~90% accuracy. These machine-learning methods do not rely on precise compound identification: Rather, the relational aspects of chromatographic and mass peaks provide the needed information, which underscores this method's utility for detecting alien biology.
Keyphrases
- machine learning
- gas chromatography
- mass spectrometry
- living cells
- big data
- artificial intelligence
- fluorescent probe
- tandem mass spectrometry
- single molecule
- liquid chromatography
- high resolution mass spectrometry
- gas chromatography mass spectrometry
- electronic health record
- ionic liquid
- high resolution
- squamous cell carcinoma
- healthcare
- risk assessment
- virtual reality
- heavy metals
- oxide nanoparticles