Identification of potential biomarkers for osteoporosis and chronic kidney disease through bioinformatics and machine learning algorithm

by Hao Tang, Kai Hu, Yuankang He, Zihao Zhang, Bingcheng Liu, Xiao Ma, Tianwen Ye Objective The purpose of this study is to identify hub genes associated with both osteoporosis (OP) and chronic kidney disease (CKD) through bioinformatics analysis, and to explore the potential pathogenetic mechanisms in OP and CKD through these hub genes. Methods We downloaded the GSE15072 and GSE56815 datasets from the GEO database as training sets, and GSE7158 and GSE70528 for validation.

Differential expression genes were selected using the “limma” package, while gene co-expression networks were constructed with “WGCNA.” Functional enrichment analyses were performed using “clusterProfiler.” Hub genes were identified through machine learning techniques, and their diagnostic efficacy was evaluated by ROC curves plotted with the ‘pROC’ package. Immune infiltration was analyzed using CIBERSORT, and pan-cancer relationships were explored to identify associations between hub genes and various tumors.

Potential therapeutic agents were investigated using the Drug Signatures Database (DSigDB). Experimental validation was conducted via RT-qPCR using cisplatin-induced chronic kidney disease (CKD) and ovariectomy (OVX)-induced osteoporosis models in C57BL/6J mice.