Skip to main content
Fig. 7 | Acta Neuropathologica Communications

Fig. 7

From: EpiDiP/NanoDiP: a versatile unsupervised machine learning edge computing platform for epigenomic tumour diagnostics

Fig. 7

Public pan-cancer UMAP plot highlighting the loss of their original epigenetic signatures. (A) Tumour cell line reference datasets cluster together (100% of cases). Note that the cell line cluster lies apart from the respective tumour of origin clusters. Axes: UMAP 0,1. (B) Subplot magnification reveals an absence of tumour entity-specific clustering. Cell lines coloured according to their origin, comprising adrenal carcinoma, urothelial carcinoma, breast cancer, cervical squamous cell cancer, colon cancer, diffuse large B cell lymphoma, endometrial cancer, oesophagal cancer, glioblastoma IDH wildtype MES/RTK I/RTK II, hepatocellular carcinoma, cutaneous melanoma, uveal melanoma, mesothelioma, lung adenocarcinoma, lung squamous cell carcinoma, ovarian cancer, pancreatic cancer, extra-adrenal paraganglioma, renal cell cancer, sarcomas, stomach cancer, germ cell tumours, renal cell cancers, and thymomas. An in-depth examination of the plot with regard to annotations and copy number profile visualisation is available at www.epidip.org and epidip.usb.ch

Back to article page