Particle.news
Download on the App Store
7 ARTICLES
·
2h ago

AI and Chemical Fingerprinting Reveal 3.3-Billion-Year Biosignatures, Recasting the Timeline of Early Photosynthesis

A peer-reviewed PNAS study uses pyrolysis–GC–MS paired with machine learning to read degraded molecular fragments as biological signals with over 90% accuracy.

Organic matter extracted from samples of 2.5-billion-year-old rock containing fossilized microorganisms like the one in this photomicrograph still contains biomolecular fragments that may have been produced via photosynthesis.
Image
Image

Overview

  • Researchers analyzed more than 400 samples spanning modern biomass, billion-year-old fossils, ancient sediments, and meteorites to train and test the model.
  • The approach extends reliable molecular detection of life from the previous ~1.7 billion-year limit to rocks dated at about 3.3 billion years.
  • Chemical patterns consistent with oxygen-producing photosynthesis were identified in rocks at least 2.5 billion years old, pushing that record back by roughly 800 million years.
  • Key detections include South African formations, with life signals from the 3.3-billion-year-old Josefsdal Chert and photosynthesis signatures from the Gamohaan Formation.
  • The team reports the method can also distinguish photosynthetic from non-photosynthetic organisms and broad cell types, and they note NASA-backed work on a rover-capable instrument alongside planned tests in Earth’s Mars-like deserts; authors stress the tool complements, not replaces, other lines of evidence.