Gene Expression Patterns Distinguish Mortality Risk in Patients with Postsurgical Shock

Figures

• 

  Figure 1. Identification of biomarker genes from gene expression data. (a) Heat map plot of genes of interest. Rows represent the gene expression value and columns represent the samples. The scale bar represents the intensity of expression of transcripts, with red indicating overexpressed transcripts and green representing underexpressed transcripts. The top bar indicates surviving (yellow) and non-surviving patients (black); (b) Kaplan–Meier plot showing survival probability of two groups of patients clustered by risk mortality. The numbers below the graph indicate the number of patients at risk of death in each group; (c) volcano plot of the differentially expressed genes, with red coloring for fold changes >1.5 and p-value < 0.01; (d) protein–protein interaction network of differentially expressed genes.

• 

  Figure 2. Relative mRNA levels of OLFM4, CD177, RETN, and IL1R2 in surviving patients and non-surviving patients as measured by qPCR. The primers and reference genes are given in the Methods section. Horizontal lines within the boxes represent the median, and the boundaries of the boxes indicate the 25th and 75th percentiles, while the whiskers indicate the highest and lowest values. The Y-axis represents the RNA expression levels in arbitrary units and logarithmic scale. qPCR, quantitative real-time polymerase chain reaction.

• 

  Figure 3. Quantification of mortality prediction accuracy by ROC AUC. (a) ROC AUC analysis of gene expression; (b) ROC AUC analysis of multivariate regression model that includes gene expression, emergency, sex, and age data, as well as creatinine, bilirubin, lactate, and white blood cell levels, as adjusted variables. ROC, receiver operating characteristic; AUC, area under the curve.

• 

  Figure 4. Survival analysis based on regression model. (a) Risk mortality tree generated by classification and regression tree (CART) analysis; (b) Kaplan–Meier curve for overall survival based on CART analysis.