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

Background Classification of large volumes of data produced in a microarray experiment allows for the extraction of important clues as to the nature of a disease. Results Using multi-dimensional unsupervised FOREL (FORmal ELement) algorithm we have re-analyzed three public datasets of skeletal muscle gene expression in connection with insulin resistance and type 2 diabetes (DM2). Our analysis revealed the major line of variation between expression profiles of normal, insulin resistant, and diabetic skeletal muscle. A cluster of most "metabolically sound" samples occupied one end of this line. The distance along this line coincided with the classic markers of diabetes risk, namely obesity and insulin resistance, but did not follow the accepted clinical diagnosis of DM2 as defined by the presence or absence of hyperglycemia. Genes implicated in this expression pattern are those controlling skeletal muscle fiber type and glycolytic metabolism. Additionally myoglobin and hemoglobin were upregulated and ribosomal genes deregulated in insulin resistant patients. Conclusion Our findings are concordant with the changes seen in skeletal muscle with altitude hypoxia. This suggests that hypoxia and shift to glycolytic metabolism may also drive insulin resistance.