Intelligent Design of Escherichia coli Terminators by Coupling Prediction and Generation Models.

Area of Science:

• Synthetic Biology
• Molecular Biology
• Computational Biology

Background:

• Transcription terminators are essential genetic regulatory elements for gene circuit design.
• Accurate characterization of terminator strength is vital for precise gene circuit engineering.
• Current computational methods for terminator strength prediction are limited by insufficient sequence and thermodynamic feature integration.

Purpose of the Study:

• To develop an intelligent design approach for *Escherichia coli* terminators.
• To construct a robust terminator strength prediction model incorporating sequence and thermodynamic features.
• To generate novel terminator sequences using deep generative models for enhanced gene circuit applications.

Main Methods:

• Extracted sequence and thermodynamic features from *E. coli* intrinsic terminators.
• Developed machine learning models for terminator strength prediction, achieving R² = 0.72.
• Employed a generative adversarial network (GAN) to learn and generate novel terminator sequences.

Main Results:

• The machine learning model demonstrated strong predictive performance for terminator strength.
• Generated terminator sequences using GANs showed similar data distributions to intrinsic terminators, validating their reliability.
• Experimental verification confirmed that 72% of selected generated terminators exhibited >90% termination efficiency.

Conclusions:

• The study successfully constructed both a terminator strength prediction model and a terminator generation model for *E. coli*.
• These models provide essential support for terminator design in synthetic biology gene circuits.
• The developed intelligent design approach enhances biological component modularity and advances synthetic biology.