MatQuantify

MatQuantify quantifies and classifies mitotic spindle and DNA features from fluorescence microscopy images to detect and quantify disruptions in spindle integrity following protein knockdown.

Key Features:

• Automated Detection and Quantification: Identifies large-scale and subtle structural changes within mitotic spindles and DNA from fluorescence microscopy images.
• Physical Property Quantification: Measures area, component lengths, perimeter, eccentricity, fractal dimension, number of satellite objects, and orientation.
• Textural Property Measurement: Computes entropy, mean intensity, and standard deviation of intensities.
• Statistical Comparison: Performs statistical comparisons to evaluate treatment effects on spindle properties.
• Machine Learning Classification: Supports training of supervised classifiers (23 algorithms tested) with a Support Vector Machine achieving 85.1% classification accuracy on spindle images.

Scientific Applications:

• Cancer Research: Analysis of mitotic spindle dynamics to study aneuploidy and identify potential therapeutic targets that induce mitotic insults in cancer cells.
• Protein Knockdown Studies: Quantitative assessment of spindle morphology and function following protein depletion experiments, including controls such as luciferase knockdown and clathrin heavy chain (CHC) depletion.

Methodology:

Analyzed fluorescence microscopy images (217 untreated metaphase, 172 luciferase knockdown, 230 CHC-depleted), quantified the listed structural and textural properties, trained 23 supervised machine learning classifiers (Support Vector Machine achieved 85.1% accuracy), and applied Kruskal–Wallis and Tukey–Kramer tests to compare properties (solidity, compactness, eccentricity, extent, mean intensity, number of satellite objects) between CHC-depleted cells and controls.