A Novel Clustering Approach: Global Optimum Search with Enhanced Positioning - Clustering Challenges in Biological Network - page 322

Fig. 16.5. Prediction of the optimal number of clusters, as shown by the turning point in the cluster
balance, by different clustering algorithms from the clustering of 5652 yeast genes based on DNA
expression levels in glucose pathway experiments. Each gene contains 36 time points, or a 24-

dimensional feature vector.

16.3.6. Coherence and Biological Relevance

Often the most intuitive and convenient manner of evaluating the robustness of
a clustering approach is to visually inspect the cluster tightness.

For brevity,

Fig. 16.6 depicts the expression time course for 4 sample clusters. We use the

Fig. 16.6. Gene expression time course plots for 4 sample clusters, found using the EP GOS Clust

algorithm.