Module Detector

One of the most important tasks in cluster analysis is to understand where the estimated clusters come from, that is, identification of the group-related features (genes). The mixed factors anslysis contains the buit-in module detector where the some sets of genes relevant to the group structure are automatically extracted. The ArrayCluster displays the expression patterns of these modules. Investigating the listed genes at these relevant modules and the visulalization would elucidate causul link from clustering to biological knowlege.


Visualizer

After running on Mixed_Factors_Analysis.exe and clicking "Relevant Module Profile(+)" / " Relevant Module Profile(-)" on the menu, the input file selection wizard will start. Selecting "relevant_set_+.txt"/ "relevant_set_-.txt" in the wizard, the expression matrix of the selected q modules will be emerged in the GUI. At a time "Select gene" window will open and ask which image of the gene expression patterns are enlarged. After selecting rows (genes) of interest, the enlarged expression image will be displayed on the right window (Figure 1).


Output File

A number of genes selected as the group-related modules are listed at "relevant_set_+.txt"/ "relevant_set_-.txt" in

The "relevant_set_+.txt" contains the q module sets where the genes in each module are selected to have the top L (user specifies at the input parametes for mixed factors analysis) of the highest positive correlation scores. On contrary, the "relevant_set_-.txt" lists q group-related modules to have negative correlation scores.
For example, if "Factor Dimension=5", "Number of Relevant Genes=20", total 100 genes and there expression values are written on each file whrer the first row denotes the file ID and the sample name, the next 20 rows denote the gene expressions of the first 20 genes to have the highest positive correlation with the first factor.


File format of relevant_set_+.txt / relevant_set_-.txt


Figure 1: Image of the relevant module profiling