卡内基跨学科团队利用随机森林模型，基于406份样本成功捕捉33亿年前生命遗迹

Core Insights - The Carnegie Institution for Science has developed a "technology fusion" solution that combines pyrolysis gas chromatography-mass spectrometry (py-GC-MS) with supervised machine learning to identify ancient biological signatures from degraded organic matter [1][2][4]. Group 1: Research Significance - The research aims to decode organic molecules buried in ancient rock layers, which is crucial for understanding Earth's history and the evolution of life [1][2]. - Traditional methods have limitations in tracing early life due to the degradation of organic remnants, making this new approach significant for filling gaps in the timeline of life's evolution [1][2]. Group 2: Methodology - The study involved a cross-disciplinary team that analyzed 406 samples, ranging from modern organisms to ancient rocks, to train the machine learning model [4][10]. - The model demonstrated a 100% accuracy in distinguishing modern organic matter from meteorite and fossil organic matter, with a 97% accuracy in identifying fossil plant tissues [2][4]. Group 3: Experimental Results - The model successfully identified biological signatures in ancient rocks dating back 3.33 billion years and 2.52 billion years, providing new methodological support for exploring earlier life traces [2][19]. - The research included a wide variety of samples, including 141 sedimentary rocks, 65 fossils, and 42 meteorites, ensuring a comprehensive dataset for model training [4][10]. Group 4: Future Implications - The fusion of py-GC-MS and machine learning represents a new paradigm in paleobiology and artificial intelligence, potentially enhancing the understanding of life's origins and the search for extraterrestrial life [24]. - The study's findings could pave the way for further innovations in identifying biological signatures in complex molecular mixtures, addressing core challenges in the field [22][24].