DeepD&Cchl: an AI tool for automated 3D single-cell chloroplast detection, counting, and cell type clustering

Area of Science:

• Plant cell biology
• Bioimaging
• Artificial intelligence in science

Background:

• Chloroplast density varies significantly across plant cell types.
• Accurate, automated single-cell chloroplast detection and quantification are essential for spatiotemporal analysis.
• Current methods lack efficiency and precision in chloroplast counting at the single-cell level.

Purpose of the Study:

• To develop an AI-powered tool, DeepD&Cchl, for automated chloroplast detection and counting in plant cells.
• To enable precise cell-type clustering based on chloroplast number and cell size.
• To provide a robust solution for chloroplast analysis across diverse imaging modalities.

Main Methods:

• Development of DeepD&Cchl utilizing the You-Only-Look-Once (YOLO) real-time detection algorithm.
• Integration of an Intersection Over Union (IOU) module for accurate chloroplast counting and error elimination.
• Combination with Cellpose for enhanced single-cell segmentation and analysis.

Main Results:

• DeepD&Cchl successfully identifies chloroplasts in plant cells across light, electron, and fluorescence microscopy images.
• The tool precisely counts chloroplasts in single- and multi-layered images, preventing double-counting.
• Cell-type clustering based on chloroplast count and cell size was effectively supported, providing morphological insights.

Conclusions:

• DeepD&Cchl represents a significant advancement in plant cell analysis, offering accuracy and efficiency.
• The AI tool facilitates precise chloroplast identification, counting, and cell-type classification.
• DeepD&Cchl serves as a valuable tool for researchers in plant biology and single-cell studies.