An AI-driven Workflow for Accelerated Optimization of Cell-Free Protein Synthesis, iScience
In a recently published article in iScience, researchers from the University of Paris-Saclay and the University of Lyon 1 have developed a software suite to optimize a cell-free system for producing small molecules.
Cell-free systems are a synthetic biology technology where biochemical reactions—such as transcription and translation—are carried out outside of living cells, in a reaction medium containing the essential components (enzymes, ribosomes, amino acids, energy). The main challenge of such system is to find the optimal proportion of each component.
The software suite developed in this study, available as a workflow on the Galaxy-SynBioCAD platform, uses the available volumes of essential components as input to initiate an active learning loop that quickly converges toward the optimal composition in just a few iterations. Apart from the active learning module, the computer code was entirely written by ChatGPT-4, with no human post-editing.
Applied to the production of various antimicrobial proteins, this software suite increased colicin production up to ninefold using cell extracts from E. coli or human HeLa cells. Activity tests confirmed that all the produced proteins were fully functional.
Ref: Khalil M, Elsawah A, Hoang A, Faulon JL, Panthu B, Hérisson J. An AI-driven Workflow for Accelerated Optimization of Cell-Free Protein Synthesis. iScience (2025). 10.1016/j.isci.2025.113599