Table 1.

Overview of the data sets, their characteristics, and the respective number of models which were built.

Dataset	Type	Number of samples	Number of replicates	Description	Number of models (consensus/sample-specific)
Data					Models
Cancer
TCGA	23 Cancer types patients	8792	mostly 1	RNA-seq data of tumour samples from various cancer types of TCGA consortium reprocessed by ref. [33]	23/8792
SKCM	Melanoma patients	472	1	RNA-seq data of SKCM tumour samples of TCGA consortium reprocessed by [33] from different stages	1/472
IN-HOUSE	Melanoma cell lines	9 cell lines (28 samples)	3–4	RNA-seq data of the A375, A375IZI, MALME 3 M, WM1346, WM1366, MeWo and SK-Mel5 melnaoma cell lines, and TUMEL patient-derived melanoma isolates	9/28
CCLE	Melanoma cell lines	49	1	RNA-seq data from 49 cell lines of CCLE collection	−/49
Control
TCGA-CONTROL	Healthy control tissues for 19 cancer types patients	740	1	RNA-seq data from the TCGA consortium of samples taken from 19 normal tissues near the tumour	19/740
TCGA-CONTROL-LIVER	Liver control patients	50	1	RNA-seq data from the TCGA consortium of liver samples near a LIHC tumour	1/50
TCGA-CONTROL-KIDNEY	Kidney control patients	129	1	RNA-seq data from the TCGA consortium of kidney samples KIRC, KIRP, KICH tumours	3/129
IN-HOUSE-CONTROL	Melanocytes	1 cell line (3 samples)	3	RNA-seq data from melanocytes	1/3

See “Material and methods” section for more details on the used cell lines and growth conditions.