ASTRA
Published:
Autonomous Multi-Agent Reconstruction of Prebiotic Reaction Networks Reveals Organizational Features of Amino Acid Synthesis
Agentic AI · Origin of Life · Reaction Networks · Amino Acid Synthesis — Preprint 2026
Decades of origin-of-life research have uncovered hundreds of prebiotically plausible reactions across hydrothermal, surface, and biochemical environments, yet this knowledge remains fragmented across chemistry, geochemistry, and astrobiology. ASTRA is a multi-agent AI system that reconstructs prebiotic reaction networks for biomolecular targets by integrating autonomous literature synthesis, reaction-network construction, and iterative mechanistic evaluation.
Applied to the twenty canonical amino acids, ASTRA reconstructed 286 reaction networks spanning 2,850 plausible pathways and 4,513 reactions — more than 92% of which are absent from existing curated databases. Across these networks it identifies recurrent organizational features (shared hubs, shared intermediate building blocks, environment-dependent molecular properties), and for literature-sparse targets such as histidine and tryptophan it generates explicit, testable bridging hypotheses.
