AI Insight
Researchers developed an autonomous AI system that independently designs, tests, and learns from protein experiments without human intervention. Operating continuously for approximately one month, multiple AI agents used robotic laboratories to explore protein variants of glycoside hydrolases, discovering enzymes with substantially altered substrate specificity toward non-native sugars while progressively mapping the sequence-function landscape. The system also uncovered unexpected determinants of substrate specificity and protein expression that were not predetermined learning objectives.
Why it matters
This represents a paradigm shift from AI that analyzes existing biological data to AI that generates new knowledge through direct experimental interaction. The autonomous experimental framework could dramatically accelerate protein engineering and biological discovery by enabling continuous, round-the-clock exploration of biological systems without requiring constant human oversight.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Biological AI learns primarily from existing observations, but many questions cannot be answered from available data alone. Here we show that AI can instead acquire knowledge by acting directly on biological systems and learning from the consequences. We developed a closed-loop framework in which autonomous agents design protein variants, construct and characterize them in a robotic laboratory, learn from the resulting experimental feedback, and decide what experiments to perform next. We then allowed the system to operate continuously and without human intervention for approximately one month, during which multiple agents independently explored protein sequence space while learning from shared experimental experience. Applied to glycoside hydrolases, the agents discovered enzymes with substantially altered substrate specificity toward non-native sugars and progressively learned the structure of the underlying sequence-function landscape. The resulting experimental experience also revealed determinants of substrate specificity and protein expression that were not specified as learning objectives. These results demonstrate that AI can autonomously interact with biology over extended periods to acquire knowledge through experience, establishing a framework for biological discovery driven by continuous experimental interaction.
Source: Learning protein function through autonomous experimental interaction