Abstract
Plant cell cultures are important biological resources widely used in various studies. The cell cultures occasionally suffer from changes in their physiological characteristics. But, little is known about genetic and epigenetic stability of the cultured cells. Moreover, an objective and quantitative method to evaluate phenotypes of the cultured cells still remains to be established. In this study, we tried to classify Arabidopsis cell lines by using their gene expression profiles. Unsupervised hierarchical clustering and principal component analysis of microarray data of 7 cell lines and AtGenExpress developmental series demonstrated that the de-differentiated cell cultures were grouped together and formed a distinct group from plant organs. The replicated samples derived from each cell line were clustered together. Similar results were obtained by self-organizing map and machine learning. These results indicate that microarray expression profiling may be useful in typing cultured cell lines.
