Figure 2

UMAP plots and confusion matrices are summarizing the classification and clustering of the cancer samples. (A, B) UMAP plots with high-degree target genes in BRCA with baseline and k-means clustering labels, respectively; (C, D) UMAP plots with PAM50 genes in BRCA with baseline and k-means clustering labels, respectively; (E, F) Confusion matrices of subtype-classification in BRCA with F1 scores with respect to the baseline labels, using high-degree target genes and PAM50 genes, respectively. Accuracy and F1 score were closer in both cases; (G) UMAP plot with all target genes using transcriptome-based baseline labels in LGG; (H) UMAP plots with high-degree target genes using expression-based baseline labels in LUSC; (I) UMAP plots with high-degree target genes using mRNA-based clusters⁸¹ as a baseline in PAAD.