Encoding Genetic Circuits with DNA Barcodes Paves the Way for Machine Learning-Assisted Metabolite Biosensor Response...

Area of Science:

• Synthetic biology
• Metabolic engineering
• Genetic circuit design

Background:

• Genetically encoded biosensors are crucial for metabolic engineering.
• Fine-tuning biosensor dose-response curves remains a significant challenge.
• Existing methods lack the capacity for large-scale biosensor characterization.

Purpose of the Study:

• To develop a novel, high-throughput workflow for constructing and characterizing genetically encoded biosensors.
• To enable the fine-tuning of biosensor dose-response curves for diverse applications.
• To facilitate the rational design of genetic circuits through comprehensive biosensor profiling.

Main Methods:

• A DNA trackable assembly method was combined with fluorescence-activated cell sorting coupled with next-generation sequencing (FACS-seq).
• A combinatorial library of 5184 FapR-fapO-based malonyl-CoA biosensor variants was constructed.
• Machine-learning algorithms were employed to predict genotype-phenotype relationships.

Main Results:

• The FACS-seq technique successfully characterized the response curves of 2632 biosensor combinations.
• Large-scale genotype-phenotype association data were generated for designed biosensors.
• A malonyl-CoA biosensor with the largest dynamic response range was successfully obtained.
• Feature importance analysis identified key design elements influencing biosensor performance.

Conclusions:

• The developed workflow provides a powerful platform for designing, tuning, and profiling biosensor response curves.
• This approach significantly advances the capabilities for creating tailored biosensors for metabolic engineering.
• The findings facilitate the rational design of complex genetic circuits with improved performance.