Statistics and Data Analysis for Microarrays Using R and Bioconductor, 2nd Edition

610 Statistics and Data Analysis for Microarrays Using R and Bioconductor

The Euclidean distance groups toge ther profiles with small differences in val-

ues re gardless whether they occur at the same or different time points. The

clusters group together profiles with expression values in the same range (e.g.,

row 1, column 1 and row 1, column 3) without distinguishing between the

different shapes within the given range. Although informative, the clustering

is not per fect: three of the resulting clusters are empty. A different initializa-

tion and a different range of pa rameters (neighborhood radius a nd learning

rate) might produce a better clustering. Given the data, the results of the

clustering are good but not exc e ptional. The SOFM was able to