Supplementary Figure 1: Sample selection procedure and TL ratio across cancer.

(a) Flowchart of sample selection. After exclusion of unsuitable samples, 18,430 samples remained. Tumor and normal samples were subsequently paired, and extra pairs per patient were dropped for a paired set consisting of n = 8,953 pairs. Further sample selection based on available data is presented in Figure 1a. The sample selection, tumor/normal pairing procedure and replicate analysis are explained in-depth in the supplementary methods. (b) Boxplot of log T/N TL ratio ‘log(Tumor TL/Control TL)’ across cancer types for all n = 8,953 samples in the paired set. Boxes colored in blue indicate that the median TL ratio for that cancer type is greater than 1 (log ratio greater than 0), and thus more than 50% of samples show TL elongation. Numbers and percentages at the top and bottom whiskers represent cancer cases with TL longer and shorter than matched normal, respectively.