NucDetect - A python package for Detection and Quantification of DNA Doublestrand Breaks

NucDetect is a Python package for the detection and quantification of γH2AX and 53BP1 foci inside nuclei. Its written in pure Python 3.13, obeys the PEP 8 style guidelines and includes PEP 484 type hints as well as Epytext docstrings.

Requirements

NucDetect is compatible with Windows, Mac OS X and Linux operating systems. It requires the following packages:

• tensorflow-cpu>=2.17.0
• scikit-image>=0.16.2
• matplotlib>=3.1.3
• seaborn>=0.13.2
• statannotations>=0.7.2
• pyqt5>=5.14.1
• numba>=0.48.0
• pillow>=11.13.0
• qtawesome==1.3.1
• piexif>=1.1.3
• pyqtgraph>=0.14.0
• pandas>=2.1.4
• imagecodecs>=2026.1.1
• openpyxl>=3.1.5
• PyWavelets>=1.9.0

Important note

While higher package versions than listed might be okay, due to the fragile nature of some of the used packages, there's a good probability that the program might break. In this case, try first to use listed minimal versions. This does not apply to the packed versions!

Installation

Run the following commands to clone and install from GitHub

$ git clone https://github.com/SilMon/NucDetect.git

or pypi

python -m pip install NucDetect

For Windows users

Download the packed program that can be found under "Releases" and place the folder where ever you desire. The program can be started by running the NucDetect.exe file.

Start

The program can be started by running the NucDetectAppQT.py:

cd %UserProfile%/AppData/local/Programs/Python/python37/Lib/site-packages/gui
python -m NucDetectAppQT

First start: Switch to the created NucDetect Folder, which will be created in User directory. Then place images you want to analyse into the images folder and click the reload button. This will load all images and create a thumbnail for each (needed to decrease the memory footprint of QT). This can take several minutes, depending on the number of images and used hardware (e.g. around 5 min for 2200 images on a Ryzen 3700X processor). Progress will be displayed in the command prompt.

Supported Image Formats

Following image formats are supported by NucDetect:

• TIFF
• PNG
• JPG
• BMP

Not supported

• Grayscale images
• Binary images

Wiki

Detailed information about the program can be found on the

Supplementary Data

https://github.com/SilMon/NucDetect_Additional_Data

---

Transparency Note

Claude Code was used during development from version 1.1.0 onwards, in the following areas:

• Code review and static analysis
• Bug and crash fixes
• Performance improvements
• Resource leak fixes (file handles, database connections)
• Development environment configuration
• Documentation

AI-assisted changes are reviewed by the authors before being merged.

Authors

Author: Romano Weiss

Co-Author: Stefan Rödiger