Toehold-VISTA: a machine learning approach to decipher programmable RNA sensor-target interactions

Area of Science:

• Synthetic biology
• Molecular diagnostics
• Computational biology

Background:

• RNA-based biosensors are crucial for synthetic biology and diagnostics.
• Designing effective RNA sensors is time-consuming and challenging.
• Understanding RNA-RNA interaction rules and structure-function relationships is limited.

Purpose of the Study:

• To develop a computational framework for rapid RNA sensor design.
• To improve the performance and efficiency of RNA biosensors.
• To accelerate RNA sensor engineering for biotechnology and diagnostics.

Main Methods:

• Developed VISTA (versatile in-silico RNA-targeting analysis), a machine learning-guided framework.
• Integrated biophysical modeling of sensor and target RNAs.
• Utilized partial least squares discriminant analysis and high-throughput experimental data for model training.

Main Results:

• VISTA successfully designs RNA sensors with enhanced performance.
• The framework captures key determinants of RNA sensor performance.
• Toehold-VISTA demonstrated improved RNA sensor design against SARS-CoV-2 RNA.

Conclusions:

• VISTA provides a broadly applicable, target-aware design strategy.
• This approach accelerates RNA sensor engineering.
• The method has significant implications for biotechnology and diagnostic applications.