Skip to main content
Fig. 4 | BMC Bioinformatics

Fig. 4

From: CancerSiamese: one-shot learning for predicting primary and metastatic tumor types unseen during model training

Fig. 4

t-SNE plots and heatmap of primary and metastatic test samples with top gene markers. a t-SNE of primary training samples with top 100 gene markers. b t-SNE of primary testing samples with top 100 gene markers. c Heatmap of the top 100 marker genes across all 29 primary cancer types. Cancers were clustered by the average per cancer type. Gene expression levels were z-transformed per gene. d t-SNE of primary and metastatic test samples together with the top 100 gene markers. e t-SNE of primary and metastatic test samples together with the top 20 gene markers

Back to article page