Evolution of Vesicle Shape Parameters Based on A Deep-Learning Based Image Analysis Framework Over Extended Time Periods

Understanding the precise analysis of shape instabilities in cell-sized membrane compartments necessitates meticulous quantification of the shape dynamics of individual components over extended timeframes. The study utilized phase-contrast microscopy to image giant unilamellar membrane vesicles, commonly employed as model systems for cellular structures.

These vesicles, composed of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine, were prepared in a fluid lamellar phase and either with or without cholesterol dissolved in a potato starch solution.

The morphological transformations and budding dynamics of individual vesicles, with varying chemical compositions, were meticulously quantified utilizing a deep-learning based image analysis framework over prolonged time periods. The input data for this automated framework consisted of phase-contrast micrographs. This sophisticated framework seamlessly integrated contour-based detection of the initial vesicle with precise segmentation of subsequent daughter compartments.

Consequently, it facilitated the explicit identification and tracking of mother and daughter vesicles at an exceptionally high temporal resolution.

These discoveries further facilitated the measurement of the temporal evolution of vesicle morphometrics, encompassing area and local membrane curvature during budding. These measurements serve as crucial inputs for developing computational models designed to elucidate curvature-driven membrane remodeling in complex soft materials.