Image analysis software design for cell culture application.

Abstract

Cell culture of mouse embryonic stem cells results in the spontaneous aggregation of these cells to form Embryoid bodies. Here they undergo differentiation spontaneously and form different cell lines. The size, spread and growth of these aggregates must be traced to understand the impact of culture media and conditions. This project aims at aiding this process, by providing a means to process and analyse the light microscopy images of the aggregates and determine their shape and size. The program developed within this project analyses the cell aggregates through a simple, effective and robust algorithm which involves segmentation, contouring, shape matching using Hu invariant moments and blob analysis. The background and methods for each of these steps is explained herein, and the algorithm is evaluated and its efficiency and accuracy determined. After the algorithm was found to be robust, it was applied to cell aggregates grown in different alginate concentrations and it was found out that an optimal concentration of 0.5-1.0% was required for stable cell aggregate sizes. This method of image analysis can be extended to other applications and cell cultures, with minor modifications and improvements.