A robust, agnostic molecular biosignature based on machine learning

• DOI: 10.1073/pnas.2307149120
• Publication Date: 2023-09-25T19:12:02Z
• Authors: H. James Cleaves; Grethe Hystad; Anirudh Prabhu; Michael L. Wong; George D. Cody; Sophia Economon; Robert M. Hazen
• Keywords: Identification

The search for definitive biosignatures—unambiguous markers of past or present life—is a central goal paleobiology and astrobiology. We used pyrolysis–gas chromatography coupled to mass spectrometry analyze chemically disparate samples, including living cells, geologically processed fossil organic material, carbon-rich meteorites, laboratory-synthesized compounds mixtures. Data from each sample were employed as training test subsets machine-learning methods, which resulted in model that can identify the biogenicity both contemporary ancient samples with ~90% accuracy. These methods do not rely on precise compound identification: Rather, relational aspects chromatographic peaks provide needed information, underscores this method’s utility detecting alien biology.