Actin Quantification Using the Filamentous Actin Segmentation Tool (FAST)

Area of Science:

• Cell Biology
• Biophysics
• Computational Biology

Background:

• Actin-filament assembly is crucial for cellular functions and disease mechanisms.
• Quantifying diverse actin structures from microscopy images is challenging due to limited tools.
• Existing methods struggle with accurate segmentation of various actin organizations.

Purpose of the Study:

• To develop and validate a deep learning model for accurate actin structure quantification.
• To introduce a novel antibody-assisted annotation method for training machine learning models.
• To provide a user-friendly tool for analyzing cytoskeletal organization in cellular imaging data.

Main Methods:

• Antibody-assisted annotation to generate labeled datasets for training.
• Development of the Filamentous actin segmentation tool (FAST) using deep learning.
• Application of FAST on phalloidin-stained and live-cell confocal microscopy images.

Main Results:

• FAST achieves accurate and efficient quantification of distinct actin structure classes.
• The antibody-assisted approach enables robust model training without structure-specific antibodies.
• Generated segmentation masks allow for quantitative analysis of actin abundance and organization.

Conclusions:

• FAST is a powerful, open-source tool for researchers studying cytoskeletal dynamics.
• The antibody-assisted annotation protocol facilitates the creation of high-quality training datasets.
• FAST simplifies the analysis of actin organization, aiding research in cell biology and disease.