GELFISH--graphical environment for labelling fluorescence in-situ hybridization images

Area of Science:

• Biomedical Imaging
• Computational Biology
• Genetics

Background:

• Automatic focusing in fluorescence in-situ hybridization (FISH) for signal counting is error-prone and time-consuming.
• A neural network approach has been developed to classify focused and unfocused signals, distinguishing valid data from artefacts.
• Accurate labeling of image signals is crucial for training this neural network classifier.

Purpose of the Study:

• To introduce GELFISH (Graphical Environment for Labelling FISH images), a novel tool for simplifying and accelerating the labeling of FISH signals.
• To enable efficient rejection of unanalyzable nuclei within FISH images.
• To provide a flexible and user-friendly environment for training classifier-based FISH signal counting systems.

Main Methods:

• Development of GELFISH, a graphical user interface implemented in popular software.
• Integration of features for simple and rapid labeling of FISH signals.
• Inclusion of functionality for the rejection of unanalyzable nuclei.

Main Results:

• GELFISH facilitates straightforward and quick labeling of FISH signals, improving accuracy.
• The system effectively identifies and allows for the rejection of unanalyzable nuclei.
• The environment is flexible, adaptable to various FISH applications, and accessible on any computer.

Conclusions:

• GELFISH significantly enhances the process of preparing data for classifier-based FISH signal counting.
• The tool streamlines image analysis, reducing time and potential errors in FISH studies.
• GELFISH offers a universally accessible solution for researchers working with FISH image analysis.