Identification and validation of biomarkers related to nicotinamide metabolic pathway activity in heart failure

Figures

• 

  FIGURE 1. Identification of NM-associated gene modules in HF. (A) The violin graph shows the scores of NM-RGs related genes in HF and control groups, with significantly higher scores in HF than in control groups. (B) The clustering diagram shows the clustering of samples. The samples are clustered according to the size of the differences between them, and the height of the line connecting two samples is the similarity degree between them. (C) Scale-free topology analysis for soft-threshold selection in WGCNA. The left figure shows the scale-free fitting index at each power value, with the red line at 0.85 indicating the threshold. The first power value above this threshold should be selected. The right figure illustrates the average number of connections at each power value. The first power value where the curve flattens after connecting the average number of connections at different power values should be chosen. Considering all factors, power = 10 is selected. (D) Hierarchical clustering dendrogram of co-expression modules. (E) Module-trait relationships showing correlation with NMRG scores. Each row in the figure represents a gene module, and the correlation between nicotinamide…

• 

  FIGURE 2. Screening of NM-related candidate genes in HF. (A) Volcano plot displaying DEGs between HF and control groups. The x-axis of the graph represents the ratio of gene expression values between patient and normal samples, with log2 applied. The y-axis shows the significance level of the gene expression difference, with-log10 applied. Red-marked genes are those that show significantly upregulated expression in the patient sample, while blue-marked genes show significantly downregulated expression. The top 10 genes with the most significant differences are highlighted on the graph. (B) Heatmap of top differentially expressed genes. Gene expression is represented by a red and blue gradient. Each row in the figure represents a gene and each column represents a sample. (C) Venn diagram showing intersection between DEGs and WGCNA module genes. (D) GO enrichment analysis of candidate genes. (E) KEGG pathway analysis of candidate genes. Each bubble in the figure represents an enrichment pathway, the color of the bubble represents the degree of significance, and the size of the bubble represents the number of genes enriched.

• 

  FIGURE 3. Identification and validation of NM-related biomarkers. (A) PPI network of candidate genes with hub nodes highlighted. (B) Core subnetwork extracted from PPI analysis. (C) LASSO coefficient profiles of candidate genes. (D) Cross-validation for optimal lambda selection in LASSO regression. The x-axis represents the logarithm of the penalty coefficient λ, and the y-axis shows the coefficients of each variable in the model. As λ increases, the coefficients of less important variables quickly drop to 0, while those of more important variables remain significant. The y-axis of Figure B represents the mean squared error (MSE), and the x-axis represents the value of λ (lambda.min). (E,F) Expression validation of candidate biomarkers in (E) training and (F) validation datasets.

• 

  FIGURE 4. Diagnostic performance evaluation of biomarkers. (A) Nomogram integrating three biomarkers for HF risk prediction. (B) Calibration curve assessing nomogram accuracy. The horizontal axis represents the predicted probability, and the vertical axis represents the actual frequency of occurrence. The ideal calibration curve is close to the 45° line, indicating accurate prediction. (C) Clinical impact curve demonstrating predictive performance.

• 

  FIGURE 5. Functional characterization of NM-related biomarkers. (A) Chromosomal localization of three biomarkers. (B) Functional similarity network among biomarkers. (C) GeneMANIA interaction network showing functional associations. (D–F) GSEA enrichment plots for (D) NDC1, (E) NUP133, and (F) TRMT11.

• 

  FIGURE 6. Immune infiltration and pathway analysis in HF. (A) Composition profile of 22 immune cell types across samples. (B) Differential immune cell infiltration between HF and control groups. (C) Correlation network between biomarkers and immune cells. (D,E) GSVA analysis of (D) Hallmark and (E) KEGG pathways showing differential activity between groups.

• 

  FIGURE 7. Regulatory networks and therapeutic prediction. (A) TF-biomarker regulatory network. (B) miRNA-mRNA interaction network. (C) Functional enrichment of key miRNAs. (D) lncRNA-miRNA-mRNA ceRNA network. (E) Drug-biomarker interaction network.

• 

  FIGURE 8. Experimental validation of biomarker expression. (A–C) RT-qPCR validation of (A) NDC1, (B) NUP133, and (C) TRMT11 expression in clinical samples from HF patients and controls.