by Juehan Wang, Haijun Li, Weiyi Shi, Xiaoxu Ren, Yingying Liu, Lin Mao, Daming Wang, Tianfang Zhang, Ziwei Zhang, Huiqin Zheng, Xiaofeng Yang, Mingfei Yao, Zuobing ChenBackground The detrimental cycle of sarcopenic obesity (SO) significantly reduces quality of life in older adults, while the mechanisms are still unclear. Materials and methods We first analyzed the incidence of SO using the CHARLS database. We identified key genes by integrating differentially expressed genes, weighted gene co-expression network analysis, and targets of gut microbiota metabolites, refining the selection through machine learning methods (LASSO, XGBoost, SVM-REF, Random Forest). These genes were validated through single-cell sequencing, receiver operating characteristic analysis, and Muscle immunohistochemistry in a high-fat-diet (HFD) induced mouse model. Further analyses comprised immune infiltration profiling, pathway enrichment, and transcriptional regulation analysis. Additionally, we explored the relationships between key genes and autophagy, ferroptosis, and immunity responses. Finally, we predicted and evaluated potential therapeutic compounds via the CMap database and molecular docking. Results SO incidence in China increased significantly from 16.1% (2011) to 20.4% (2018). Machine learning identified ALDH1A3, CSF1R, and PHGDH as key genes. These genes were validated in external muscle single-cell datasets, demonstrating robust diagnostic performance with AUC values exceeding 0.72 across four independent GEO cohorts. Following an HFD intervention in mice, ALDH1A3 and CSF1R expression in muscle tissue was significantly upregulated, while PHGDH showed a consistent upward trend that did not reach statistical significance. Immune infiltration analysis revealed a significant increase in resting NK cells in both obesity and sarcopenia states. Functional enrichment analyses using Gene Ontology and Kyoto Encyclopedia of Genes and Genomes linked the genes to transcriptional regulation pathways. The Cisbp_M4923 motif was identified as the most relevant transcription factor binding site. Finally, molecular docking simulations indicated stable binding of the top candidate compound, Birinapant, to the key gene targets. Conclusion ALDH1A3, CSF1R, and PHGDH serve as potential co-morbid biomarkers for SO.