Novel AI Technique Reveals Sample’s Biological or Non-Biological Origin with 90% Accuracy

The search for definitive biosignatures — unambiguous markers of past or present life — is a central goal of paleobiology and astrobiology.

A team of researchers led by the Carnegie Institution for Science has developed a robust method that combines pyrolysis gas-chromatography mass-spectrometry (GC-MS) measurements of a wide variety of terrestrial and extraterrestrial carbonaceous materials with machine-learning-based classification to achieve 90% accuracy in the differentiation between samples of abiotic origins vs. biotic specimens, including highly-degraded, ancient, biologically-derived samples.

Article
@EverythingScience