Related Experiment Video
Updated: Aug 30, 2025
Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
Development and Verification of a Combined Diagnostic Model for Sarcopenia with Random Forest and Artificial Neural
Shangjin Lin1,2, Cong Chen1, Xiaoxi Cai1
1Department of Orthopeadic, Huadong Hospital Affiliated to Fudan University, Shanghai 200040, China.
Background:
Sarcopenia is a chronic disease characterized by an age-related decline in skeletal muscle mass and function, and diagnosis is challenging owing to the lack of a clear "gold standard" assessment method.
Objective:
This study is aimed at combining random forest (RF) and artificial neural network (ANN) methods to screen key potential biomarkers and establish an early sarcopenia diagnostic model.
Methods:
Three gene expression datasets were downloaded and merged by searching the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs) in the merged dataset were identified by R software and subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. Afterward, the STRING database was employed for interaction analysis of the differentially encoded proteins. Then, RF was used to identify key genes from the DEGs, and a sarcopenia diagnostic model was constructed by ANN. Finally, the diagnostic model was assessed using a validation dataset, while its diagnostic performance was evaluated by the area under curve (AUC) value.
Results:
107 sarcopenia-related DEGs were identified, and they were mainly enriched in the FoxO and AMPK signaling pathways involved in the molecular pathogenesis of sarcopenia. Thereafter, seven key genes (MT1X, FAM171A1, ZNF415, ARHGAP36, CISD1, ETNPPL, and WISP2) were identified by the RF classifier. The proteins encoded by three of these genes (CISD1, ETNPPL, and WISP2) may be potential biomarkers for sarcopenia. Finally, a diagnostic model for sarcopenia was successfully designed by ANN, achieving an AUC of 0.999 and 0.85 in the training and testing datasets, respectively.
Conclusion:
We identified several potential genetic biomarkers and successfully developed an early predictive model with high diagnostic performance for sarcopenia. Moreover, our results provide a valuable reference for the early diagnosis and screening of sarcopenia in the future.
