Deep-learning deconvolution and segmentation of fluorescent membranes for high-precision bacterial cell-size...

Area of Science:

• Microbiology
• Evolutionary Biology
• Cell Biology

Background:

• Bacterial evolutionary studies often overlook cell-size variation.
• Existing methods for cell-size measurement have limitations.

Purpose of the Study:

• Introduce MEDUSSA, a high-throughput method for precise bacterial cell-size profiling.
• Investigate cell-size diversity in bacterial populations.
• Identify genetic factors influencing cell size.

Main Methods:

• Utilized deep-learning-based membrane deconvolution and segmentation for fluorescent images.
• Developed error-corrected cell measurement for accurate single-cell dimensions.
• Applied the MEDUSSA method to six strains of Priestia megaterium.

Main Results:

• MEDUSSA provides accurate cell dimensions, overcoming limitations of phase-contrast methods.
• Observed over twofold differences in cell volume across Priestia megaterium strains.
• Identified a partially-functional PBP1 allele associated with reduced cell width in one strain.

Conclusions:

• MEDUSSA is a powerful tool for studying bacterial cell-size diversity.
• Comparative analyses are crucial for understanding bacterial cell biology.
• The study expands the toolkit for investigating the evolution of bacterial cell size.