To check how well our list of differentially expressed genes distinguishes one treatment group from another, we can perform hierarchical clustering based on the gene list. Clustering can also be used to discover novel groups within your data set, identify gene expression signatures distinguishing groups of samples, and to identify genes with similar patterns of gene expression across samples. 

• Select the Feature list data node
• Select the Visualizations section of the task menu
• Select Hierarchical clustering from the Visualizations section of the task menu (Figure 1)
  
  

The Hierarchical clustering menu will open (Figure 2).

• Select Finish to run with default settings

A Hierarchical clustering task node will be added to the pipeline (Figure 3).

• Double-click the Hierarchical clustering task node to launch the heat map

The Dendrogram view will open showing a heat map with the hierarchical clustering results (Figure 4).

Samples are shown on rows and genes on columns. Clustering for samples and genes is shown through the dendrogram trees. More similar samples/genes are separated by fewer branch points of the dendrogram tree. 

The heat map displays standardized expression values with a mean of zero and standard deviation of one. 

The heat map can be customized to improve data visualization using the menu on the left-hand side of the heat map. 

• Select Treatment from the Attributes drop-down menu

Samples are now labeled with their Treatment group (Figure 5). 

Samples cluster based on treatment group and the 5μM and 10μM groups are more similar to each other than to the 0μM group. 

We can save the heat map as a publication-quality image. 

• Select 
• Choose size and resolution using the Save as SVG dialog (Figure 6)
  
  

• Select Save

The heat map will be saved as a .svg file and downloaded in your web browser. 

For more information about hierarchical clustering and the Dendrogram view, please see the Hierarchical Clustering user guide.