Segmentation and classification of cell cycle phases in fluorescence imaging

Area of Science:

• Cellular Biology
• Biochemistry
• Microscopy Imaging

Background:

• Studying the correlation between biochemical networks and cell cycle progression in live cells often relies on fluorescence imaging of fusion proteins.
• Fluorescently tagged proteins like GFP-PCNA generate dynamic sub-cellular patterns that indicate cell cycle phases, but image analysis is challenging due to variable patterns and noise.

Purpose of the Study:

• To develop a sophisticated and efficient algorithm for automatic nucleus segmentation and cell cycle classification from live-cell fluorescence microscopy data.
• To improve the scalability of existing methods for high-content imaging analysis.

Main Methods:

• Extension of graph partitioning active contours (GPAC) for nucleus segmentation, incorporating regional density functions for enhanced accuracy.
• Utilizing surface shape properties of the GFP-PCNA intensity field to derive descriptors for foci patterns.
• Automated cell cycle phase classification based on these derived descriptors.

Main Results:

• Dramatically improved efficiency of the GPAC algorithm, making it scalable for high-content microscopy.
• Successful automated cell cycle phase classification using GFP-PCNA foci patterns.
• Quantitative performance evaluation demonstrated accuracy comparable to manually labeled data.

Conclusions:

• The enhanced GPAC algorithm provides a robust and efficient solution for analyzing live-cell fluorescence microscopy data.
• This method facilitates reliable automated segmentation and cell cycle classification, crucial for understanding biochemical networks and cell cycle dynamics.