Identification of Diagnostic Biomarkers for Fibromyalgia Using Gene Expression Analysis and Machine Learning

Fuyu Zhao¹Jianan Zhao¹Yang Li²Chenyang Song¹Yaxin Cheng¹Yunshen Li¹Shiya Wu²Bingheng He^3*Juan Jiao^2*Cen Chang^1*

• ¹Department of Rheumatology, Shanghai Guanghua Hospital, Shanghai University of Traditional Chinese Medicine, Shanghai, China
• ²Guang’anmen Hospital, China Academy of Chinese Medical Sciences, Beijing, Beijing Municipality, China
• ³Tongren Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China

Objective: Fibromyalgia (FM) is a complex autoimmune disorder characterized by widespread pain and fatigue, with significant diagnostic challenges due to the absence of specific biomarkers. This study aims to identify and validate potential genetic markers for FM to facilitate earlier diagnosis and intervention.Methods: We analyzed gene expression data from the Gene Expression Omnibus (GEO) to identify differentially expressed genes (DEGs) associated with FM. Comprehensive enrichment analyses, including GO, KEGG, and Reactome pathways, were performed to elucidate the biological functions and disease associations of the candidate genes. We employed the eXtreme Gradient Boosting (XGBoost) algorithm to develop a diagnostic model, which was validated using independent datasets.Results: Three genes, Dual Specificity Tyrosine Phosphorylation Regulated Kinase 3 (DYRK3), Regulator Of G Protein Signaling 17 (RGS17), and Rho Guanine Nucleotide Exchange Factor 37 (ARHGEF37), were identified as key biomarkers for FM. These genes are implicated in critical processes such as ion homeostasis, cell signaling, and neurobiological functions, which are perturbed in FM. The diagnostic model demonstrated robust performance, with an Area Under the Curve (AUC) of 0.8338 in the training set and 0.8178 in the validation set, indicating its potential utility in clinical settings.Conclusion: The study successfully identifies three diagnostic biomarkers for FM, supported by both bioinformatics analysis and machine learning models. These findings could significantly improve the diagnostic accuracy for FM, leading to better patient management and treatment outcomes.