Tutorial: Transcriptome Clustering

This brief tutorial uses a cancer cell microarray expression dataset to demonstrate AutoSOME transcriptome clustering.

1) If you have not already done so, Launch AutoSOME.

2) Download and save the example cancer line expression dataset. (primary dataset: Alizadeh et al. (2000) Nature, 403:503)

3) Select the Input button to launch a file browser. Load the example dataset.

4) After loading the dataset, expand Basic Fields and change the Cluster Analysis field to columns. AutoSOME is now configured to cluster transcriptomes rather than individual gene probes. Next, expand Input Adjustment and select Unit Variance normalization.

5) Since AutoSOME clusters transcriptomes by first generating an all-against-all similarity matrix, the GUI provides three commonly used distance metrics for comparing transcriptomes to one another. Expand Advanced Fields and locate the Fuzzy Cluster Networks panel. Although Euclidean distance is selected by default, one can also use Pearson's or Uncentered correlation metrics (Let's keep the default for this tutorial). See FAQ for an overview of the distance metrics.

6) Run AutoSOME with all remaining parameters set to their default values (i.e. No. Ensemble Runs = 50; P-Value <= 0.1).

7) All clusters in the output tree are shown below using the heat map display. We see that the heat map shows a clustered similarity matrix of cell lines (shown here using a contrast-adjusted rainbow heat map).

8) The necessary output files to create Fuzzy Cluster Networks (Edges and Nodes text files) are now available and can be viewed by selecting them from the Output Files table.

9) This concludes the mini-tutorial on transcriptome clustering.