ArrayStar

Heat Map


The Heat Map view uses color to display expression levels of many genes across a number of experiments. The colors of the tiles represent the measured expression value of a given gene in a given experiment.


A three color scale is used to illustrate expression level differences; default colors are:

  • blue indicating low expression values,
  • yellow color indicating intermediately expressed genes,
  • red color representing highly expressed genes
  • colors can easily be changed to suit the users preference.
  • heatmap


    Gene Names are displayed to the right of each corresponding row. As rows become smaller and closer together, the gene names are displayed in a list for each position of the mouse. The actual gene your mouse is hovering over will be underlined in the abbreviated list, as shown above.


    A Gene Tree is provided on the left of the Heat Map. This shows the clustering of different genes as calculated by the selected clustering method. Clicking on a node will illustrate a sub-tree of the genes. Clustering of sub-trees can also be performed.


    The Heat Map window also includes an Info Pane that provides additional information.

    A.        A Histogram of the data distribution in the Heat Map is displayed above the color scale. As selections are made on the Heat Map, the selected genes are shown in gray within the histogram.

    histogram


    The Histogram features include:

  • Minimum and maximum expression levels
  • The ability to use either absolute or relative color scales
  • Identification (in gray) and location of selected genes

  • B.        The Overview section displays the Clustering Type that was used for the current data set, as well as the Distance Metric and Linkage Method, if applicable. Also shown is the number of data points plotted in the Heat Map, the number of tree nodes, and the maximum tree depth.



    Distance Metrics that ArrayStar can employ include:
  • Euclidean
  • Euclidean-Squared
  • Standard Pearson's correlation
  • Uncentered Pearson's correlation
  • Manhattan (city block)
  • Linkage methods are used in Hierarchial clustering to define how data points are clustered, based on the distance between them. Methods available for ArrayStar users include:

  • Single linkage
  • Complete linkage
  • Centroid linkage
  • Average linkage

  • C.        The Details section, as mentioned above, displays information about the genes currently selected in the Heat Map. If no genes are selected, simply the number of genes in the clustering is displayed. If only one gene is selected, available details for that gene are shown, including the name, annotations, and notes. In cases where multiple genes are selected, the number of genes selected is shown.



    In addition, as the cursor moves over the Heat Map, details for each gene are displayed including
  • the gene name,
  • annotations,
  • notes,
  • the expression level for the experiment the mouse is hovering over.

  • D.        The Actions section of the Info Pane provides an easy guide to different follow-up actions that the user may want to pursue including:
  • Saving the clustering results
  • Remembering the selection set as a gene set
  • Showing the table of selected genes
  • View the current clustering results in a Line Graph Thumbnail