Gene Expression Analysis Using Fuzzy ART

The recent advances of genome-scale sequencing and array technologies have made it possible to monitor simultaneously the expression pattern of thousands or tens of thousands of genes. One of the following steps is to discover or extract the information for the genetic networks by analyzing such massive data sets. Therefore, various clustering methods, such as hierarchical clustering [3] or selforganizing maps [5], have been examined and used to elucidate the fundamental or/and characteristic expression pattern. We have applied a fuzzy adaptive resonance theory (Fuzzy ART) model, a type of unsupervised clustering method, to the experimental data [6]. In the present paper, we verified the clustering results using Fuzzy ART by comparing with those of hierarchical clustering, k-mean clustering and self-organizing maps (SOMs).