Integrative Transcriptomic Analysis Identifies Common Biomarkers in Type 1 and Type 2 Diabetes Using WGCNA and Machine Learning
The origins of type 1 (T1D) and type 2 diabetes (T2D) are etiologically distinct; however, their clinical manifestations suggest overlapping molecular mechanisms. We applied an integrative transcriptomic strategy to identify common biomarkers in diabetes subtypes. Using GSE9006 peripheral blood mononuclear cells (PBMC) expression data from Affymetrix U133A/B platforms, we integrated weighted gene co-expression network analysis (WGCNA) and differential expression analysis (DEGs). We initially identified disease-associated gene modules (T1D: 78–126 genes; p < 0.01) and DEGs with |log₂FC| > 0.5 and adjusted p < 0.05. By intersecting the module genes and DEGs, three consistently upregulated genes shared by both subtypes: MXD1, NAMPT, and KCNJ15. These genes are involved in apoptosis regulation, immune-metabolic signaling, and insulin secretion, indicating a shared molecular basis in diabetes pathology. Their biomarker potential was validated through machine learning classification, which achieved highly predictive performance for both diabetes types with area curve (AUC) 0.857–0.875. The results reveal conserved molecular mechanisms in diabetes disease pathology and demonstrate the strength of transcriptomic in identifying robust biomarkers across clinically distinct but mechanistically overlapping disorders.
Authors:
Muhammad Rezki Rasyak, Turyadi, Rudi Nirwantono, Fitya Syarifa Mozar, Alyssa Imani, Advendio Desandros, Bens Pardamean
The 6th International Conference on Biosciences (ICoBio) 2025