MitoSegNet

MitoSegNet segments mitochondria in fluorescence microscopy images using a pretrained deep learning framework to quantify mitochondrial morphology.

Key Features:

• Deep Learning Segmentation Model: Built on a pretrained deep learning framework tailored for mitochondrial segmentation, performing pixelwise and morphological segmentation tasks.
• Performance Superiority: Demonstrated superior performance in comparative studies versus traditional feature-based segmentation algorithms and Ilastik, yielding precise segmentations across experimental conditions.
• Versatility Across Samples: Applied to fluorescence microscopy images including mitoGFP-expressing mitochondria in wild-type and catp-6^ATP13A2 mutant C. elegans adults and to HeLa cells treated with fragmentation-inducing reagents.
• Morphological Analysis Integration: Integrates morphological analysis to quantify mitochondrial shape and organization from segmentation results.

Scientific Applications:

• Mitochondrial morphology quantification: Enables quantitative analysis of mitochondrial shape and network organization to study mitochondrial function and dynamics.
• Model organism studies: Supports comparative morphological analysis in C. elegans using mitoGFP and catp-6^ATP13A2 mutant samples.
• Mammalian cell studies: Supports segmentation and morphological assessment of HeLa cells subjected to fragmentation-inducing reagents.

Methodology:

Uses a pretrained deep learning segmentation model trained on diverse microscopy datasets to recognize and segment mitochondrial structures and perform pixelwise and morphological segmentation.