Genetic Optimizer Fine-Tunes Cell Cultures in Real Time

A controller design for genetic systems that overcomes the context-dependent limitations of other such optimizers has been developed by researchers at New York University Abu Dhabi (NYUAD), the University of Florida, and the University of California, Berkeley. By combining elements of both synthetic biology and genetically encoded biosensors, it adapts to changes in the cell culture environment in real time, allowing rational, robust fine-tuning of cellular processes.

The system develops genetic modules that “locate and track implicit reference signals to create synthetic circuits that can sense and respond to environmental changes, and then adjust their activity for optimal performance,” according to the authors of a recent paper in Nature Communications.

“In cell-based manufacturing, the optimum conditions are often unknown, time-varying, and sensitive to environmental changes,” Andras Gyorgy, PhD, assistant professor electrical engineering and bioengineering, NYUAD, and first author, tells GEN. “The optimizer module ensures this unknown optimum is automatically located and tracked despite environmental perturbations and changes, and without any external intervention. This occurs in a pathway-agnostic fashion, independently of the particular pathway that needs to be regulated.”

The controller “steers the regulated genetic system towards the optimum, but…without explicit knowledge of its location, the objective function…or the pathway that needs to be regulated,” the researchers wrote.

More specifically, the optimizer adjusts the production and decay rate of a set of regulator species. In the study, using parameters that are typical in Escherichia coli, the authors showed that it could locate and track the optimum based on mass action kinetics dynamics. The implementation is underpinned by the interconnection of logic gates, bistable switches, oscillators, and other common biomolecular sensors and actuators. Likewise, it relies on genetic components that already are available: “protein-based transcription controls, inducible degradation via Mesoplasma florum Lon protease and ssrA tag, a repressilator-based oscillator, CRISPRi-based toggle switches, and STAR-based logic gates.”

Closed loop performance can adjust to considerable parameter variations in multiple contexts. This technology can be implemented for minimum/maximum seeking and multi-species optimization. It also can help maintain optimal growth rates despite cellular stress.

“It can be easily integrated with existing pathways and genetically encoded biosensors to ensure its successful deployment in a variety of settings,” the paper notes. The next step is to integrate the modules and test the system.

Applications range from metabolic engineering to sustainable biomanufacturing. Limitations are the lag time between changes in the regulator and its impact on the reporter, and shifts that occur faster than the oscillator period. Therefore, this method may be best-suited to cell cultures in which the environmental shifts occur over many hours.

The MATLAB code to implement and deploy this optimizer are publicly available.

Previous articleIncreasing Machine Learning Use in Bioprocessing       
Next articleCross-Reactive Antibodies Link Multiple Sclerosis to Common Virus