GelGenie: an AI-powered framework for gel electrophoresis image analysis

Area of Science:

• Biochemistry
• Bioinformatics
• Laboratory Automation

Background:

• Gel electrophoresis is a standard technique for biomolecule separation and analysis.
• Traditional gel image analysis methods have seen limited advancement, lagging behind AI applications in other scientific domains.
• Existing software for gel electrophoresis analysis lacks ease-of-use and versatility.

Purpose of the Study:

• To develop and validate an AI-based system for automated gel band identification.
• To improve the speed, accuracy, and versatility of gel electrophoresis image analysis.
• To provide an accessible, open-source tool for researchers to analyze gel electrophoresis data.

Main Methods:

• Training various U-Nets using a dataset of over 500 manually labeled gel images.
• Employing image segmentation to classify pixels as 'band' or 'background' for precise band identification.
• Developing GelGenie, an open-source application for on-device gel image analysis.

Main Results:

• The AI system accurately identifies gel bands across diverse experimental conditions in seconds.
• The AI system demonstrates superior ease-of-use and versatility compared to current gel analysis software.
• Quantitative results from the AI system on external datasets match original author findings.
• GelGenie provides a user-friendly platform for automated gel band extraction without requiring expert knowledge.

Conclusions:

• AI-powered segmentation offers a significant advancement in gel electrophoresis image analysis.
• The developed AI system and GelGenie application enhance the efficiency and accessibility of biomolecular analysis.
• This approach has the potential to standardize and improve quantitative analysis across various gel electrophoresis applications.