Supplementary Information           (BEST VIEWED WITH NETSCAPE)

 

Comparisons and validation of statistical clustering techniques for microarray gene expression data

 

Susmita Datta and Somnath Datta

 

Bioinformatics 19, 2003

 

 

 

Simulated data sets: 

Expression profiles for all the genes in the simulated datasets  and

the nine model temporal profiles

 

 

 

 

The average proportion of non-overlap measure for various clustering algorithms

applied to the simulated data

 

 

 

Average temporal profiles of nine groups obtained using various 

clustering algorithmsfor the simulated datasets

 

Sporulation Data:   The results of the grouping are graphically represented below. Here the points represent the first two principal components of the actual data (for the positively expressed genes). The points are color-coded and labeled by their group membership which is local to each plot (since a group label under one clustering algorithm does not correspond to a group label under another algorithm). These graphs should only be compared visually to one another.

 

Clustering of all positively expressed genes during sporulation

 

 

 

Hand-picked genes (ORF) in various temporal groups

 

 

 

 

Expression profiles of all positively expressed genes.

The clusters are obtained using Diana

 

S-Plus code used in the paper to compute the validation measures for the Sporulation data  Download

 

 


FastCounter by bCentral

 

 

since  December 1, 2003