Development and validation of a gene signature predicting the risk of postmenopausal osteoporosis

Abstract

Aims: We aimed to develop a gene signature that predicts the occurrence of postmenopausal osteoporosis (PMOP) by studying its genetic mechanism.

Methods: Five datasets were obtained from the Gene Expression Omnibus database. Unsupervised consensus cluster analysis was used to determine new PMOP subtypes. To determine the central genes and the core modules related to PMOP, the weighted gene co-expression network analysis (WCGNA) was applied. Gene Ontology enrichment analysis was used to explore the biological processes underlying key genes. Logistic regression univariate analysis was used to screen for statistically significant variables. Two algorithms were used to select important PMOP-related genes. A logistic regression model was used to construct the PMOP-related gene profile. The receiver operating characteristic area under the curve, Harrell's concordance index, a calibration chart, and decision curve analysis were used to characterize PMOP-related genes. Then, quantitative real-time polymerase chain reaction (qRT-PCR) was used to verify the expression of the PMOP-related genes in the gene signature.

Results: We identified three PMOP-related subtypes and four core modules. The muscle system process, muscle contraction, and actin filament-based movement were more active in the hub genes. We obtained five feature genes related to PMOP. Our analysis verified that the gene signature had good predictive power and applicability. The outcomes of the GSE56815 cohort were found to be consistent with the results of the earlier studies. qRT-PCR results showed that RAB2A and FYCO1 were amplified in clinical samples.

Conclusion: The PMOP-related gene signature we developed and verified can accurately predict the risk of PMOP in patients. These results can elucidate the molecular mechanism of RAB2A and FYCO1 underlying PMOP, and yield new and improved treatment strategies, ultimately helping PMOP monitoring.Cite this article: Bone Joint Res 2022;11(8):548-560.

Keywords: Gene signature; Postmenopausal osteoporosis; Quantitative real-time polymerase chain reaction; Weighted gene co-expression network analysis; actin; biomarkers; calcium; decision curve analysis; gene expressions; logistic regression analysis; macrophages; univariate analysis.

Fig. 1

Identification of postmenopausal osteoporosis (PMOP)-related subtypes. a) A consistent cumulative distribution function (CDF) graph and a δ area graph with different cluster numbers (k = 2 to 9). b) The consensus matrix heat map yielded the identification of three clusters. c) Principal component analysis (PC1 and PC2) shows evident clusters among the three PMOP subtypes. d) to f) Gene set variation analysis (GSVA) enrichment analysis shows the biological pathways between different PMOP subtypes. The red colour represents the activated pathway, and the blue colour represents the inhibited pathway. d) PMOP cluster A and PMOP cluster B; e) PMOP cluster A and PMOP cluster C; f) PMOP cluster B and PMOP cluster C.

Fig. 2

Identification of the core modules related to postmenopausal osteoporosis (PMOP). a) Analysis of the scale-free index of various soft threshold powers. b) Analysis of the mean connectivity of various soft threshold powers. c) Dendrogram of all differentially expressed genes clustered based on dissimilarity measurements (1-TOM). d) Heat map of the characteristic genes of the module. We chose the blue module for subsequent analysis. ME, module eigengene; TOM, topological overlap matrix.

Fig. 3

Functional enrichment analysis of postmenopausal osteoporosis (PMOP) hub genes. a) Biological process analysis. b) Cell component analysis. c) Molecular function analysis.

Fig. 4

The selection of postmenopausal osteoporosis (PMOP) feature genes by two algorithms: a) Least Absolute Contraction and Selection Operator (LASSO) algorithm; b) support vector machine-recursive feature elimination (SVM-RFE) algorithm; c) the intersection of characteristic genes screened by the two algorithms. CV accuracy, cross-validation accuracy.

Fig. 5

Construction and verification of a postmenopausal osteoporosis (PMOP)-related gene signature. a) The best parameter (lambda) in the least absolute contraction and selection operator (LASSO) model is selected to pass the minimum criterion using five-fold cross-validation. b) Distribution map of LASSO coefficients of five features. c) The nomogram of PMOP gene characteristics. d) The receiver operating characteristic (ROC) curve is used to evaluate the prognostic value of the gene signature. e) The calibration curve of the nomogram. f) Decision curve analysis of the nomogram. g) The ROC curve of the validation group verifies that the gene signature have good prognostic value. AUC, area under the curve; CI, confidence interval; FPR, false positive rate – the proportion of the number of negative samples misjudged as positive in the actual negative samples; TPR, true positive rate – the proportion of predicted correct positive samples to all actual positive samples.

Fig. 6

Correlation analysis between genes for constructing gene signatures and immune cells. Red represents positive correlation; blue represents negative correlation. *p < 0.05; **p < 0.01.

Fig. 7

The gene expression of the gene signature in clinical samples. a) Member RAS oncogene family (RAB2A) was amplified in seven samples, except for sample 3; b) FYVE and coiled-coil domain autophagy adaptor 1 (FYCO1) was amplified in seven samples, except for sample 3.

Fig. 8

Gene set enrichment analysis (GSEA) in high- and low-risk group samples. a) Samples of the high-risk group were enriched in the calcium signalling pathway, differentiated cardiology pathway, neuroactive live receiver interaction, olfactory production, retinol metabolism, and other pathways. b) Samples in the low-risk group were enriched in myocardial contraction, citrate cycle, dilated cardiomyopathy, adhesive spot, hypertrophic cardiomyopathy, and other pathways.