Fig. 1

The whole workflow chart of the study. Seven processes include as follows: ① Pan-cancer data collection from TCGA, ② Performing differential gene expression analysis (mutated versus wild type samples) for each cancer related genes annotated in COSMIC database, ③ Identification of 105 onco-signatures, each one composed by specific marker genes, ④ Performing of the single-sample Mann–Whitney–Wilcoxon gene set test in TCGA-BRCA dataset, ⑤ Identification of the 28 most prognostic breast cancer onco-signatures, by applying the Cox proportional hazard model, ⑥ Performing the clustering analysis on the matrix of normalized enrichment scores calculated through the single-sample Mann–Whitney–Wilcoxon gene set test, showing on the rows the 28 breast survival associated onco-signatures and on the columns the TCGA primary breast cancer samples, ⑦ Molecular characterization of the identified groups through survival analysis, differential expression analysis, copy number variation analysis, DNA methylation analysis, promoter analysis, gene ontology enrichment analysis