Unraveling the Molecular Heterogeneity in type 2 Diabetes: A Potential Subtype Discovery Study Followed by Metabolic Modeling

Abstract

Background Type 2 diabetes mellitus (T2DM) is a complex multifactorial disease with a high prevalence in the world. Insulin resistance and impaired insulin secretion are the two major abnormalities in the pathogenesis of T2DM. Skeletal muscle is responsible for over 75% of the glucose uptake, thus plays a critical role in T2DM. Here, we attempted to provide a better understanding of abnormalities in this tissue. Methods We have explored the muscle gene expression pattern in healthy and newly diagnosed T2DM individuals using supervised and unsupervised classification along with examining the potential of sub-typing based on the gene expression pattern in patients.Results A machine-learning technique applied to identify a pattern of gene expression that could potentially discriminate between normoglycemic and diabetic groups. A gene set comprises 26 genes identified, which was able to discriminate healthy from diabetic individuals with 94% accuracy after 10-fold stratified cross-validation. In addition, three distinct clusters with different dysregulated genes and metabolic pathways identified in diabetic patients. Conclusion This study implies that it seems the disease has triggered through different cellular/molecular mechanisms and it has the potential to be sub-typing. Possibly, subtyping of T2DM patients in combination with their real clinical profiles will provide a better understanding of abnormalities in each group and lead to the recommendation of the appropriate precision therapy for each subtype in the future.