Figure 3: Transcriptomic and proteomic clustering of breast cancer PDX and human samples.

(a) Transcriptomic clustering of PDX breast tumours based on the PAM50 gene expression markers. (b) Proteomic clustering of PDX breast tumours based on the top 436 variably expressed proteins. (c) Phosphoproteomic clustering of PDX breast tumours based on the top 1,737 variably expressed phosphosites. (d) Proteomic clustering using only 133 non-differential expressed proteins between WHIM and human breast tumour samples. The clustering reproduced the basal-enriched and luminal-enriched clusters, where PDX (n=24) and TCGA human breast tumour samples (n=77) cluster based on their subtypes. The non-differentially expressed proteins were identified through a t-test with FDR>0.3 between the PDX and the TCGA human tumour samples. The PDX tumours are labelled by their WHIM IDs whereas the human tumours are not labelled.