(13) Cluster genes
At this point, you can generate a series of clusters using four different methods. Clustering is a very popular process for DNA microarrays, so we will describe this first, but remember that exploration is equally valid and may tell you more about your genes and experimental conditions than clustering can. Exploring your data can be performed any time after segmentation. All you need to explore are expression files (*.exp).
With MAGIC Tool, there are four ways to cluster genes. You can cluster from any dissimilarity file. First you have to calculate the clusters and then you can display them in a variety of ways. The most common way to cluster is called hierarchical clustering, which you can do with MAGIC. However, we prefer Q-T clustering (see Instructor's Guide for details), but Hierarchical Clustering is the only format currently compatible with the data visualization program Java TreeView. You can also cluster by k-means or supervised clustering.
Once you have clustered the genes, you can display the results in several ways. MAGIC allows you to view these clusters in a variety of dynamic displays. Each display can be saved as an image file for publishing or teaching. Display options are addressed in more detail later in this manual.