Skip to main content
Figure 1 | BMC Genomics

Figure 1

From: Unsupervised clustering of gene expression data points at hypoxia as possible trigger for metabolic syndrome

Figure 1

FOREL Clustering Partially Separates Normal from DM2 and IGT. A. The set of 43 samples (Mootha et al. data set) from the skeletal muscle of normal glucose tolerance (NGT), impaired glucose tolerance (IGT) and diabetic (DM2) breaks into a series of clusters spread along a line in the gene expression space. Each cluster is represented by a sphere with the radius equal to the distance from the centroid to the most distant member of the cluster. The color is proportional to the percentage of diabetic samples in the cluster, ranging from red (100% normal) to blue (DM2). Cluster 8 and cluster 1 represent singletons (i.e. single subjects that do not cluster with the other 41 samples). The percentage of diabetes in each cluster varied along principal component 2. B. Analysis of the prevalence of NGT and IGT across clusters revealed that the number of subjects with NGT dropped as distance from Cluster 3 increased. Similarly, the number of subjects with IGT increased as distance from Cluster 3 increased. The two clusters with the greatest distance from the 'normal' cluster 3, Cluster 6 and Cluster 7, 1/2 were diagnosed with DM2 and all other members of these clusters have IGT. Principle Components are presented as the value × 104.

Back to article page