An automatic image analysis approach to quantify stained cell cultures

Abstract

Counting cells in culture is a common task in biotechnology research and production. This process should be automated to provide fast and objective quantification. Flow cytometry is adapted to count cells in suspension. However, the morphological information and the spatial organisation of adherent cells are lost when cells are removed from culture. This paper proposes a methodology based on image analysis to quantify stained nuclei in culture. The protocol is composed of several steps: cell staining, automatic microscopy imaging, segmentation by an automatic algorithm including a classification approach, and computation of quantitative data that characterizes the growth of cells. An evaluation shows that the automatic process of counting provides results similar to human manual counting. The major interests of the proposed approach are the fully automated processing and preservation of cell shapes and positions in culture. More than two thousand culture conditions have been measured by this tool for various applications including optimization of cell culture media, improvement of the culture processes and measurement of drug toxicity.