Construction of a 5-feature gene model by support vector machine for classifying osteoporosis samples

Area of Science:

• Genomics
• Bioinformatics
• Immunology

Background:

• Osteoporosis is a prevalent bone disease in the elderly.
• Current classification methods for osteoporosis patients are insufficient.
• Accurate patient classification is crucial for effective treatment and management.

Purpose of the Study:

• To construct a gene signature for accurate prediction and classification of osteoporosis patients.
• To identify key genes involved in osteoporosis development.
• To develop a reliable diagnostic tool for osteoporosis.

Main Methods:

• Acquired and analyzed three Gene Expression Omnibus (GEO) datasets for osteoporosis.
• Screened differentially expressed genes (DEGs) using Limma package in R.
• Constructed a protein-protein interaction (PPI) network for DEGs.
• Developed a Support Vector Machine (SVM) classifier based on feature genes.
• Performed pathway enrichment analysis using clusterProfiler.

Main Results:

• Identified 310 DEGs associated with immune responses and protein secretion.
• Established a PPI network comprising 12 key DEGs.
• Developed an SVM classifier using five feature genes with high prediction accuracy and AUC.
• Validated the classifier across multiple osteoporosis datasets (GSE35959, GSE62402, GSE13850, GSE56814, GSE56815, GSE7429).

Conclusions:

• A robust SVM classifier based on a five-gene signature accurately predicts and classifies osteoporosis.
• The identified genes represent potential biomarkers for osteoporosis.
• This gene signature may aid in personalized medicine approaches for osteoporosis management.