Automated Image Analysis for Quantification of Protein Localization in Budding Yeast

M. Kvarnström, K. Logg, J. Hagmar, K. Bodvard and M. Käll. The 9th Int. Conf. on Systems Biology, Gothenburg, Sweden, August 2008.

Objective: Key insights in protein function can be provided by studying the sub-cellular and temporal localization of proteins in vivo and fluorescence microscopy provides an excellent tool to conduct such measurements. However, due to the subjective and qualitative nature of human interpretation of images, manual objective and unbiased  analysis of images is not straightforward and for large-data sets it becomes practically impossible.

Results: We present fast image analysis algorithms for automated cell recognition in bright field images of populations of budding yeast cells. The algorithms do not rely on fluorescent staining of the cell membrane and they are therefore in particular suitable for in vivo studies. After the cell contour has been determined, assessment of various cell morphology parameters is straightforward. Spatial analysis of the corresponding fluorescence image is also possible and includes e.g. determination and classification of protein localizations. We exemplify this by identifying protein abundance to three different types of spatial configurations corresponding to cell nucleus, plasma membrane and peroxisomes.

Conclusions: The cell recognition method is fast and robust against variations in experimental parameters such as clustering of cells. It also adapts well to variations in cell density, and illumination level. This makes the algorithms suitable for large scale studies where the major part of the analysis has to be done without human intervention. The performance of cell recognition and contour extraction was tested on more than 1000 cells in 25 images and it was found that 96% of the cells were correctly defined. We also demonstrate how the shape criterion used here for budding yeast cells can be adapted to other types of cell shapes such as fission yeast and red blood cells.




  • Mats Kvarnström, Fraunhofer-Chalmers Centre
  • Jonas Hagmar, Fraunhofer-Chalmers Centre
  • Katarina Logg, Dept. of Applied Physics, Chalmers University of Technology
  • Kristofer Bodvard, Dept. of Applied Physics, Chalmers University of Technology
  • Mikael Käll, Dept. of Applied Physics, Chalmers University of Technology



Photo credits: Nic McPhee