Automated Classification of Cellular Phenotypes Using Machine Learning in Cellprofiler and CellProfiler Analyst

Area of Science:

• Cell biology
• Computational biology
• Biophysics

Background:

• Cellular images contain vast phenotypic data imperceptible to the human eye.
• Automated image analysis and machine learning offer unbiased methods for studying cellular mechanisms and pathology.

Purpose of the Study:

• To develop a customized image analysis pipeline for detecting and quantifying changes in E-Cadherin localization and adherens junction morphology.
• To leverage CellProfiler and CellProfiler Analyst for advanced cellular image analysis.

Main Methods:

• Utilized CellProfiler for generating image-based measurements of cellular features.
• Employed CellProfiler Analyst's machine learning capabilities for data analysis.
• Focused on quantifying E-Cadherin localization and adherens junction morphology.

Main Results:

• Successfully detected and quantified alterations in E-Cadherin localization.
• Characterized changes in the morphology of adherens junctions.
• Enabled unbiased analysis of cellular phenotypes.

Conclusions:

• The developed pipeline provides a robust method for analyzing complex cellular phenotypes.
• Automated analysis enhances the understanding of cellular mechanisms and disease-related effects.
• This approach facilitates high-throughput, quantitative assessment of cell junction dynamics.