Back

		Data Analysis We are currently using primarily two clustering packages to analyze our  microarray data. The first is Cluster by Michael Eisen.      This package will let the investigator analyze multiple datasets simultaneously      (e.g. a time course study), filter the dataset (e.g. use only data > 3 SD), adjust the      dataset (e.g. normalize) and cluster the dataset. The clustering options in the program      are hierarchical clustering, K-mean clustering, Self-Organizing Maps, and principle      component analysis. Each of these methods have parameters useful to set a threshold of      significance. The tab-delimited text file output of Cluster can be read by another      companion program called Treeview. This program is an interactive graphical analysis      package designed to view data generated by Cluster. The GUI interface presents the data      in a tree format next to a color representation of the clustered array data and allows the      investigator to quickly zoom in and identify the genes in specific clusters. The second is Genecluster by Golub et al..      This computer program will create self-organizing maps (SOM) using gene expression      data. Additionally, the package will let one filter and normalize the data across multiple      array experiments as well as visualize the clusters. Alternative commercial packages are also being evaluated.
		Workflow We are in the process of creating web based workflow schemes to manage the array workflow. Visit again soon for more information!