CascadeMAP: Autonomous Microfluidic Self-driving Lab for Optimizing Enzyme Cascades

Résumé fourni par la source

Enzyme cascades demand laborious optimization of multiple parameters, such as enzyme ratios, pH, buffer composition or temperature, requiring efficient data collection and learning hidden patterns in the data. Here, we introduce CascadeMAP, an autonomous microfluidic self-driving lab for optimizing enzyme cascades. CascadeMAP presents a synergistic fusion of microfluidics for rapid collection of high-fidelity data at a minimal cost for reagents and machine learning to intelligently navigate the experiments [1]. CascadeMAP was employed to optimize two model multienzyme systems: (i) a glycerol detection cascade and (ii) a metabolic pathway for degradation of toxic pollutant 1,2,3-trichloropropane [2]. Those two examples showcase both fluorescence and label-free Raman detection. By employing the Bayesian optimization [3], we demonstrate a fully autonomous convergence to optimal parameters within a single experiment consisting of 100 consecutive learning cycles conducted completely without human intervention [4]. Thanks to its versatility, we anticipate that CascadeMAP will emerge as a valuable tool in metabolic engineering for optimizing complex metabolic networks.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.