Fig. 1

Unsupervised hierarchical clustering based on proportions of mutational signatures in each tumour. a Unsupervised hierarchical clustering on substitution and rearrangement signatures revealing four main clusters. b Clinical and mutational annotation for each sample. c Heatmap of the normalised contribution of substitution and rearrangement signatures identified in the cohort. d BRCAness predictions: HRDetect prediction score (scores above 0.7 considered BRCAness) and RNA classifier predictions. e Somatic substitution and indel driver mutations are present in more than three samples. f Tumour mutational burden of somatic mutations per Mb in the coding region. g Number of substitutions fitted to mutational signatures for each tumour. h Number of rearrangements fitted to mutational signatures for each tumour