Gene expression changes To quantify the expression level of each gene in each sample we used the dataset from Wang and colleagues74 where RNA-seq data of both tumour and control samples from TCGA—alon

Data source: The authors used a dataset from Wang et al. which contains RNA-seq data from both tumor and control samples from TCGA, as well as expression data from normal samples from the GTEx consortium.

Data normalization: The data was quantile-normalized and batch-corrected using ComBat77.

Quantification of gene expression: The expression level of each gene in each sample was quantified.

Differential expression analysis: Differential expression was computed as a log2 fold change between expression in cancer versus a matched normal sample, and then averaged across samples.

Handling missing values: If the expression of a gene was not measured in either the normal or the matched cancer type-specific samples, then the expression omic value for that gene was not computed and its missing value was set to zero