Mixed factors analysis contains a set of visualization tools for high-dimensional microarray data. This task can be addressed by calculating factor scores in much the same way as the classical factor analysis. A radical difference of our method from the classical one is that the compressed data constructed in the mixed factors analysis explicitly reflects the existing group structure of data, while the classical method ignores the presence of groups during the data compression process.


Visualizer

After running Mixed_Factors_Analysis.exe and clicking "Factor Score Matrix", the wizard to visualize the box plot matrix of factor scores will start. There, user can select three styles of sample names that are labeled at data points. By viewing "Plot Image" in "Plot Style Configuration", user can select one option. Then press "Finish".

Figure 1 displays a factor score matrix constructed with factor dimension equal to 7. By investigating each data plot, user can understand the existing groups of data, and elucidate a casual link between each coordinate of factor variable and biological explanation.

The ArrayCluster can use zoom function of Lunascape. By the scale up or down, user can focus on a part of box plots of interest. Or, by going to "Factor Score" menu and selecting two axes of interest in the "Select Mixed Factor Wizard", a large scaled box plot can be displayed.







Output File

A data file for plotting factor scores, mixed_factors.txt is created at

• \C:ArrayCluster\extents\lunamacplugin\data\ .

A part of data in mixed_factors.txt are displayed in below. The first column contains the sample names. The following seven columns correspond to the computed scores of the tissue samples for the seven ;coordinates of factor vector.


File format of mixed_factors.txt



Figure 1: Box plot of the computed factor scores