Result of hierarchical clustering of features

The result of a feature clustering is shown in figure 25.40.

Image feature_clustering_result
Figure 25.40: Hierarchical clustering of features.

If you have used an experiment (Image experiment) as input, the clustering is added to the experiment and will be saved when you save the experiment. It can be viewed by clicking the Show Heat Map (Image heatmap) button at the bottom of the view (see figure 25.41).

Image experiment_editors
Figure 25.41: Showing the hierarchical clustering of an experiment.

If you have selected a number of samples ( (Image array) or (Image rnaseq)) as input, a new element will be created that has to be saved separately.

Regardless of the input, a hierarchical tree view with associated heatmap is produced (figure 25.40). In the heatmap each row corresponds to a feature and each column to a sample. The color in the $ i$'th row and $ j$'th column reflects the expression level of feature $ i$ in sample $ j$ (the color scale can be set in the side panel). The order of the rows in the heatmap are determined by the hierarchical clustering. If you place the mouse on one of the rows, you will see the name of the corresponding feature to the left. The order of the columns (that is, samples) is determined by their input order or (if defined) experimental grouping. The names of the samples are listed at the top of the heatmap and the samples are organized into groups.

There are a number of options to change the appearance of the heat map. At the top of the Side Panel, you find the Heat map preference group (see figure 25.42).

Image heatmap_side_panel
Figure 25.42: Side Panel of heat map.

At the top, there is information about the heat map currently displayed. The information regards type of clustering, expression value used together with distance and linkage information. If you have performed more than one clustering, you can choose between the resulting heat maps in a drop-down box (see figure 25.43).

Image heatmap-dropdown
Figure 25.43: When more than one clustering has been performed, there will be a list of heat maps to choose from.

Note that if you perform an identical clustering, the existing heat map will simply be replaced. Below this box, there is a number of settings for displaying the heat map.

Below you find the Samples and Features groups. They contain options to show names, legend, and tree above or below the heat map. Note that for clustering of samples, you find the tree options in the Samples group, and for clustering of features, you find the tree options in the Features group. With the tree options, you can also control the Tree size, from tiny to very large, and the option of showing the full tree, no matter how much space it will use.

For clustering of features, the Features group has an option to "Optimize tree layout". This attempts to reorder the features, consistently with the tree, such that the most expressed features form a diagonal from the top-left to the bottom-right of the heat map.

The Samples group contains an "Order by:" dropdown that allows re-ordering of the columns of the heat map. When clustering by samples it is possible to choose between using the "Tree" to determine the sample ordering, and showing the "Samples" in the order they were input to the tool. When clustering by features, only the "Samples" input order is available.

Note that if you wish to use the same settings next time you open a heat map, you need to save the settings of the Side Panel.