What Does Superposition of Expression Data Mean?
Within the framework of SVD, the expression of each gene and array is a superposition (or a weighted sum) of all of the eigengenes and eigenarrays, respectively. You could think of the SVD’s mathematical separation of the expression patterns of the genes and arrays into eigengenes and corresponding eigenarrays, respectively, as attempting to unravel the overall expression signal into its generating components: independent experimental and biological processes, and the corresponding cellular states. In other words, SVD can be used to attempt to describe expression data as the outcome of a simple network, where a few independent sources of expression, experimental or biological, affect all the genes in the dataset (Figure 10). Classification of genes and arrays using hierarchical clustering results in groups of genes of similar contributions from all the different sources of expression. You may be interested, however, in grouping genes according to the effects of only a subset of the dif
