Fifty one yeast genes related to DNA repair and DNA recombination from microarray expression data were analyzed to construct the gene network. In this study, we consider the microarray data sets from two independent synchronized methods: Alpha factor and cdc28. Hierarchical clustering algorithm and K-means are applied to cluster the 51 genes. The clustering pictures of the 51 genes, we show that Hierarchical clustering out performs the k-means. Multiple lag correlations are need as the criterion to construct the gene networks of these 51 genes, among which 16 genes are extrapolated. The gene networks of the 16 genes is useful to find out activator or repressor genes. In addition, the 16 genes are clustered into the same subcluster by Hierarchical clustering method. The results of our studies provide a useful economic and time saving experiment procedure.

Diss. (sammanfattning) Stockholm : Karolinska institutet, 2006. / Härtill 6 uppsatser.