Related Experiment Video
Updated: Jul 13, 2026
08:13
Manipulation and Analysis of Cell Cycle-Dependent Processes in Budding Yeast
Published on: September 26, 2025
Automated image analysis of protein localization in budding yeast
Shann-Ching Chen1, Ting Zhao, Geoffrey J Gordon
1Center for Bioimage Informatics, Department of Biomedical Engineering, Carnegie Mellon University, Pittsburgh, PA 15213, USA.
Bioinformatics (Oxford, England)
|July 25, 2007
Summary
We developed an automated system to classify yeast protein locations, achieving 81% agreement with previous manual assignments. This computational method offers objective and repeatable protein localization for Saccharomyces cerevisiae research.
Area of Science:
- Molecular and Cellular Biology
- Bioinformatics
- Yeast Genetics
Background:
- Saccharomyces cerevisiae, the first sequenced eukaryote, has extensive gene and protein annotations.
- Protein subcellular localization is crucial for understanding protein function.
- The UCSF yeast GFP fusion localization database manually classified protein locations for 75% of yeast proteins into 22 categories.
Purpose of the Study:
- To develop computational methods for automated analysis of yeast protein subcellular localization.
- To create a system that recognizes the same 22 location categories used in the UCSF study.
- To provide an objective, quantitative, and repeatable method for assigning protein locations in yeast.
Main Methods:
- Developed computational algorithms to analyze yeast GFP fusion images.
- Trained the automated system on existing manually classified yeast protein localization data.
- Applied the system to 2640 yeast GFP images.
Main Results:
- The automated system achieved 81% agreement with previous manual classifications across 2640 images.
- High confidence assignments showed 94.7% agreement.
- Visual inspection suggested potential for improved accuracy in automated assignments.
- The method does not require colocalization with marker proteins.
Conclusions:
- Automated analysis of yeast protein localization is feasible and accurate.
- The developed system provides objective, quantitative, and repeatable protein localization data.
- This method can be applied to new yeast image datasets, including different strains or conditions.
