Development and validation of risk prediction and neural network models for dilated cardiomyopathy based on WGCNA

Figures

• 

  Figure 1. Implementation of WGCNA analysis and acquisition of common genes. (A) The gene co-expression module under the gene tree is represented using different colors. (B) Heatmap depicts the association between modules and DCM. The numbers inside and outside of the brackets represent p values and correlation coefficients, respectively. (C) A correlation map between the gene significance and module membership in the blue module. (D) Venn diagram of common genes acquisition. DEGs, differentially expressed genes; DCM, dilated cardiomyopathy; WGCNA, Weighted Gene co-expression network analysis; IGs, interesting genes.

• 

  Figure 2. Analysis of gene enrichment pathways and construction of PPI networks. (A) Gene enrichment analysis of CGs. The color green denotes CC, the color red denotes BP, the color purple denotes MF, and the color blue denotes the KEGG pathway. (B) Upregulated genes are indicated by nodes with red circles, whereas downregulated genes are indicated by nodes with green triangles. CGs, common genes; CC, cellular component; BP, biological process; MF, molecular function; GO, gene ontology; and KEGG, Kyoto encyclopedia of genes and genomes.

• 

  Figure 3. Validation of the hub genes in the DCM mice model. (A–H) Relative expression of Aspn, Podn, Htra1, Mfap4, Fap, Ltbp2, Fndc1, and Islr in DCM mice. DCM group, n = 5; control group, n = 5. Data are presented as the mean ± standard deviation. DOX, doxorubicin (***represents p < 0.001; **represents p < 0.01; *represents p < 0.05; and ns represents p > 0.05).

• 

  Figure 4. Construction and validation of nomogram. (A) Development of a prediction nomogram based on validated hub genes. (B) Calibration curves for predicting the risk of dilated cardiomyopathy nomogram. (C) ROC curve analysis of risk prediction model in train group. (D) ROC curve analysis of risk prediction model in the test group.

• 

  Figure 5. Construction and validation of neural network model. (A) Result from neural network visualization. (B) ROC curve of the neural network model in train group (GSE57338). (C) ROC curve of the neural network model in the test group (GSE116250).

• 

  Figure 6. Differential expression analysis of experimentally validated genes in the test group. Expression levels of genes (A) ASPN, (B) FAP, (C) HTRA1, (D) PODN, and (E) MFAP4 in normal and DCM samples. The expression of these genes was found to be significantly upregulated in DCM samples compared to normal samples (****represents p < 0.0001; ***represents p < 0.001).

• 

  Figure 7. Diagnostic efficiency assessment of validated hub genes in the test group. ROC curve analysis of (A) FAP, (B) MFAP4, (C) PODN, (D) ASPN, and (E) HTRA1 in the test group.

• 

  Figure 8. Analysis of immune cell infiltration in patients with DCM. (A) Differences in immune microenvironment between healthy and DCM samples. (B) Heatmap depicting the correlation between five hub genes and 23 immune cells. Correlation between (C) ASPN, (D) HTRA1, and infiltrating immune cells in DCM and healthy samples.