Construction of GSE6631- and TCGA-HNSCC-specific modules using WGCNA. (A) Network topology for different soft-thresholding powers of GSE6631. Left: the scale-free topology fit index (y-axis) in the GSE6631 dataset; right: the mean connectivity (y-axis) as a function of the soft-thresholding power (x-axis) in the GSE6631 dataset. (B) Network topology for different soft-thresholding powers of TCGA-HNSCC. Left: the scale-free topology fit index (y-axis) in the TCGA-HNSCC dataset; right: the mean connectivity (y-axis) as a function of the soft-thresholding power (x-axis) in the TCGA-HNSCC dataset. (C) Branches of the clustering dendrogram from different modules based on differentially expressed genes (DEGs) in GSE6631. Based on the unsupervised hierarchical clustering, GSE6631 DEGs were clustered into seven modules (yellow, blue, green, red, brown, turquoise, and gray). (D) Branches of the clustering dendrogram from different modules based on DEGs in TCGA-HNSCC. Based on the unsupervised hierarchical clustering, TCGA-HNSCC DEGs were clustered into nine modules (black, red, pink, blue, brown, turquoise, green, yellow, and gray). (E) Module-clinical feature relationships in the GSE6631 modules. In the heatmap, red represents a positive correlation, whereas blue represents a negative correlation. The numbers in each block are the related-value and P-value of the correlation. (F) Module-clinical feature relationships in TCGA-HNSCC modules. In the heatmap, red represents a positive correlation, and blue represents a negative correlation. The numbers in each block are the related-value and P-value of the correlation.