Clustering gene expression data from colon tissue samples with an interrelated two-way self-organizing map
Abstract
We use an Interrelated Two-Way Self-Organizing Map (ITW-SOM) to cluster gene expression data from normal and tumor colon tissue samples. Results show 3 optimal tissue sample clusters; two of which belong to the normal and tumor samples, and a third intermediary cluster, which comprises of tissue samples whose features could be ascribed as an intermediate between normal and tumor states. The Davies-Bouldin index provided the means to assess the number of optimal clusters for both genes and tissue samples.