ImageJ Analysis of PTBP1 Localization in Heterokaryon Assays

This project investigates the intracellular localization of PTBP1 (Polypyrimidine Tract Binding Protein 1) and its RRM2 domain in mammalian cells. Using ImageJ (Fiji) and custom Python pipelines, I quantified nuclear vs. cytoplasmic localization from fluorescence microscopy images.

---

🔬 Project Goals

• Analyze fluorescence microscopy images from heterokaryon assays
• Quantify PTBP1 subcellular distribution with automated image analysis
• Determine whether the RRM2 domain mediates nuclear retention
• Develop reproducible computational workflows for biological image analysis

---

🧰 Tools & Technologies

• ImageJ/Fiji (custom segmentation and quantification workflows)
• Python (OpenCV, scikit-image, pandas for batch processing)
• Statistical Analysis (scipy, statsmodels for hypothesis testing)
• Data Visualization (matplotlib, seaborn, Plotly for results presentation)
• Interactive Dashboard (Dash for data exploration)

---

🎥 Live Demo

Interactive Dash dashboard showing real-time exploration of PTBP1 localization data. Users can filter by experimental conditions, visualize N/C ratios, and generate publication-ready plots.

---

📂 Repository Structure

ptbp1-imagej-analysis/
│── data/
│   ├── raw/BBBC013/images/          # Raw microscopy images
│   ├── processed/                   # Analysis results
│   └── raw_images/                  # Plate map metadata
│── scripts/
│   └── analyze_ptbp1.py            # Batch image processing
│── notebooks/
│   └── analysis.ipynb              # Statistical analysis
│── app/
│   └── app.py                      # Interactive dashboard
│── figures/
│   ├── localization_plots/         # Visualization outputs
│   └── gif/                        # GIF demonstrations
│── requirements.txt                # Python dependencies
└── README.md

---

📊 Example Output

Example visualization showing nuclear/cytoplasmic ratio distributions across experimental conditions

• Automated analysis of 26,040 cells from 96 microscopy images
• N/C ratio quantification for protein localization patterns
• Statistical comparisons between experimental conditions
• Initial results were presented at CSUF Summer Research Symposium (2021) for this project

---

🔮 Future Work

• Integrate deep learning segmentation (StarDist/Cellpose)
• Develop cloud-based analysis pipeline
• Expand to multi-protein co-localization studies
• Create web interface for collaborative research

---

Note: This pipeline uses the BBBC013 public dataset for demonstration purposes. Results validate the analytical approach rather than represent specific biological findings.