filoVision - using deep learning and tip markers to automate filopodia analysis

Area of Science:

• Cell Biology
• Biotechnology
• Microscopy

Background:

• Filopodia are crucial cellular structures for environmental sensing.
• Analyzing filopodia typically requires specific markers (actin, tip, or membrane).
• Existing methods pose challenges for diverse cell types and visualization preferences.

Purpose of the Study:

• To develop an automated filopodia analysis workflow.
• To address the need for analyzing amoeboid filopodia using only tip labels.
• To create a flexible deep learning platform for filopodia research.

Main Methods:

• Development of the filoVision deep learning platform.
• Implementation of filoTips for labeling filopodia tips and cytosol.
• Integration of filoSkeleton for analyzing filopodia shafts with actin labeling.
• Utilizing the ZeroCostDL4Mic framework for adaptability.

Main Results:

• FiloVision enables automated analysis of filopodia using single tip markers.
• FiloTips facilitates information extraction without actin or membrane markers.
• FiloSkeleton provides comprehensive analysis of filopodia shafts and tip proteins.
• The platform is adaptable to various cell types and user datasets.

Conclusions:

• FiloVision offers a flexible and automated solution for filopodia analysis.
• The platform enhances accessibility for researchers studying cell exploration.
• FiloVision supports diverse visualization methods and cell types in filopodia research.