Exploring Key Genes with Diagnostic Value for Nonalcoholic Steatohepatitis Based on Bioinformatics Analysis

Figures

• 

  Figure 1. Sample typing of NASH patients based on the GSE48452 dataset. (A) Volcano map of differential analysis of the GSE48452 dataset. The red dots indicate upregulated genes, green dots indicate downregulated genes, and black dots indicate genes without significant expression (control, n = 11; NASH, n = 17). (B) Consistency clustering matrix of k = 2. (C) Consistency clustering matrix of k = 3. (D) Consensus cumulative distribution map. Consensus cluster CDF from k = 2 to k = 10. (E) Delta area plot from k = 2 to k = 10. Relative change of the area under the CDF curve. (F) Cluster–consensus plot from k = 2 to k = 10. (G) Comparison of the fibrosis index and NAS between sample typing; cluster 1, n = 6; cluster 2, n = 6; cluster 3, and n = 14. *p < 0.05; **p < 0.01.

• 

  Figure 2. GSEA results of comparison between specific genes for each sample and normal samples. (A) Enrichment plot of comparison between cluster 1 specific genes and normal samples. (B) Enrichment plot of comparison between cluster 2 specific genes and normal samples. (C) Enrichment plot of comparison between cluster 3-specific genes and normal samples. The black vertical line represents the specific genes in the sample typing; the abscissa indicates the difference between the sample typing and the normal samples. The genes closer to the left are the genes with a more significant difference.

• 

  Figure 3. Modules associated with the clinical parameters of NASH based on WGCNA. (A) Cluster dendrogram of liver tissue samples from 26 NASH patients. (B) Scale-free fitting index (left) and average connectivity (right) of various soft threshold power β. The red line indicates the correlation coefficient (0.9). (C) Cluster dendrogram of the coexpressed genes. Each leaf on the cluster dendrogram corresponds to a different gene module. (D) Heat map of correlation between modules and patient clinical parameters (Fat, NAS, and fibrosis index). Each cell contains the corresponding correlation and P-values. (E) Heat map of the expression of the red, blue, brown, and yellow module genes in the normal samples, cluster 1, cluster 2, and cluster 3 samples.

• 

  Figure 4. Protein interaction network of 162 sample genotyping-specific genes. (A) Protein interaction network of 162 sample genotyping-specific genes. (B) Ranking diagram of degree values (top 20). (C) Bubble plot of KEGG pathway enrichment results for the top 20 genes; the horizontal coordinate indicates GeneRatio, the vertical coordinate indicates the pathway entry name, the bubble color indicates P adj, and the bubble size indicates counts.

• 

  Figure 5. Risk model construction and clinical correlation analysis. (A) Heat map of DEGs in the training group. (B) Abscissa represents the log(λ) value, the ordinate represents the binomial deviance, the upper part is the number of genes retained by the corresponding log(λ) value for the calculation, and the dashed line indicates the log(λ) value and the number of genes. (C) ROC analysis of the constructed risk model and the NASH occurrence in the train and test groups. (D) ROC analysis of the risk model constructed and NASH occurrence in all samples. (E) Correlation between the risk model value and fibrosis index. (F) Correlation between the risk model value and NAS.

• 

  Figure 6. Diagnostic value of five sample typing-specific genes and correlation with lipid metabolism. (A–E) Correlation between the risk model values with SCD (A), ACLY (B), PNPLA2 (C), PGC1α (PPARGC1A) (D), and LPL (E) expression.