Fig. 12

WGCNA based on the gene expression spectrum of the dataset. A Sample clustering was used to detect anomalous samples; each branch represents 1 sample and the red box in the sample information indicates the category of each sample. B Scale-free topological model fit (left) and mean connectivity (right), used to determine the best soft threshold. The horizontal coordinate of the left panel indicates the soft threshold of fit and the vertical coordinate indicates the level of fit (R-squared) in the scale-free topological model. The right panel indicates the soft threshold of fit in horizontal coordinates and the average connectivity between modules in vertical coordinates. C Dynamic shear clustering tree for different genes. The upper tree shows gene co-expression, with each gene represented by a leaf in the tree and each module represented by a main trunk branch. The lower colored bars indicate the corresponding 17 modules, labelled in the indicated colors. D Heatmap of the correlation between different modules and the sepsis group; x-axis refers to sample classification and y-axis to the module color. Red in the correlation heatmap indicates positive correlation and dark blue indicates negative correlation. Textual meanings of the color blocks in the heatmap: Pearson correlation coefficient in the upper part; statistical p-value in the lower brackets. E Turquoise module genes and mitochondrial autophagy differential genes intersection