ParticleAnalysis: A Python Tool for Single-Particle Tracking in Fluorescence Microscopy

Overview

ParticleAnalysis is an open-source Python application for detecting, tracking, and analyzing fluorescent particles in microscopy image sequences. It provides a user-friendly graphical interface for analyzing single-particle dynamics, with particular emphasis on robust particle detection, reliable tracking, and comprehensive motion analysis.

Key Features

• Advanced Particle Detection

  • Multi-scale Gaussian detection for varying particle sizes
  • Local background subtraction and noise estimation
  • SNR-based filtering of detections
  • Interactive ROI selection and parameter tuning

• Robust Particle Tracking

  • Nearest-neighbor linking algorithm with gap closing
  • Support for particle disappearance and reappearance
  • Track quality control and filtering
  • Visualization of tracking results

• Comprehensive Motion Analysis

  • Mean Square Displacement (MSD) analysis
  • Diffusion coefficient calculation
  • Motion type classification (confined, normal, directed, anomalous)
  • Track shape analysis (radius of gyration, asymmetry, fractal dimension)

• User-Friendly Interface

  • Interactive visualization of particles and tracks
  • Real-time parameter adjustment
  • Results tables with sorting and filtering
  • Batch processing capabilities

Installation

# Create a new conda environment
conda create -n particle_analysis python=3.11
conda activate particle_analysis

# Install dependencies
pip install -r requirements.txt

# Install the package
pip install .

Quick Start

1. Launch the application:

python -m particle_analysis

2. Load data:

   • Click "Open File" to load a TIFF stack
   • Adjust contrast using the histogram sliders
   • Select a region of interest (optional)

3. Detect and track particles:

   • Adjust detection parameters if needed
   • Click "Detect Particles" to identify particles
   • Click "Track Particles" to create trajectories
   • Or use "Analyze Full Image" for complete analysis

4. Analyze results:

   • View tracks in the Visualization tab
   • Examine feature distributions in the Analysis tab
   • Export results using File → Save Results

Usage Examples

# Programmatic usage
from particle_analysis.core import ParticleDetector, ParticleTracker
from particle_analysis.io import DataReader

# Load data
reader = DataReader()
movie = reader.read_movie('example.tif')

# Detect particles
detector = ParticleDetector(min_sigma=1.0, max_sigma=3.0)
particles = detector.detect_movie(movie)

# Track particles
tracker = ParticleTracker(max_distance=5.0, max_gap=2)
tracks = tracker.track_particles(particles)

Technical Details

Particle Detection

The detection algorithm employs a multi-scale Gaussian fitting approach:

1. Background estimation using median filtering
2. Local maxima detection with SNR thresholding
3. Sub-pixel localization through 2D Gaussian fitting
4. Quality control based on fit parameters

Tracking Algorithm

Particle linking is performed using a nearest-neighbor approach with the following steps:

1. Frame-to-frame particle assignment using KDTree
2. Gap closing for temporary disappearances
3. Track initialization and termination
4. Track quality filtering

Motion Analysis

Track analysis includes:

• MSD calculation and curve fitting
• Diffusion coefficient estimation
• Anomalous diffusion exponent calculation
• Shape descriptors computation
• Motion type classification

Data Format

The application accepts:

• Input: TIFF stacks (.tif, .tiff)
• Output: CSV files containing:
  • Particle positions and intensities
  • Track coordinates and features
  • Analysis results

Requirements

• Python 3.11 or later
• PyQt6
• NumPy
• SciPy
• pandas
• scikit-image
• pyqtgraph

Contributing

Contributions are welcome! Please see our Contributing Guidelines for details.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• The scikit-image team for image processing algorithms
• The trackpy project for inspiration on particle tracking approaches
• The PyQt team for the GUI framework

Contact

For questions and support:

• Open an issue on GitHub
• Email: george.dickinson@gmail.com