Phenotypic switching in gene regulatory networks

Area of Science:

• Systems Biology
• Molecular Biology
• Genetics

Background:

• Gene expression noise can cause reversible phenotypic switching in cells.
• Protein abundance distributions in isogenic cell populations often show multiple peaks, indicating distinct phenotypes.
• Quantifying these multimodal distributions is crucial for understanding cellular decision-making.

Purpose of the Study:

• To develop a methodology for quantifying multimodal gene expression distributions and single-cell power spectra in gene regulatory networks.
• To provide a practical analytical tool for studying complex nonlinear gene regulatory networks.

Main Methods:

• Extending the linear noise approximation (LNA).
• Approximating multimodal distributions as a mixture of Gaussian components in the limit of slow promoter dynamics.
• Analyzing gene regulatory networks, including two-promoter networks and genetic oscillators.

Main Results:

• A closed-form approximation for multimodal gene expression distributions was derived.
• The method was applied to various genetic networks, revealing new dynamical characteristics of phenotypic switching.
• The interplay of transcriptional and translational regulation was shown to control gene expression distribution multimodality.
• Phenotypic switching was linked to bimodal expression in genetic oscillators and hysteresis in phenotypic induction.

Conclusions:

• The developed methodology offers a practical tool for analyzing complex gene regulatory networks beyond stochastic simulation.
• Gene regulatory networks possess inherent memory capabilities, demonstrated through phenotypic induction hysteresis and bimodal expression.
• Understanding noise-driven phenotypic switching is key to deciphering cellular decision-making processes.