Fig. 4

The landscape of non-ref NUMTs in 663 ESCC patients. a The number of non-ref NUMTs per sample. The dashed lines indicate the average number of non-ref NUMTs per sample type. b The number of non-ref NUMTs of the germline, loss, and somatic per individual. c The correlation between the frequency and size of the 110 distinct non-ref NUMTs detected in all 1326 samples, with summary histograms at the edges. The high-frequent non-ref NUMTs may include more than one type of source indicated by different colors and shapes. The distribution of non-ref NUMTs with frequency < 0.2 and size < 2500 bp was zoomed in. d The circos plot of 110 distinct non-ref NUMTs. From the outside: (1) mtDNA genes (left) and nDNA chromosome (right); (2) frequencies of non-ref NUMTs reported in Wei W. et al. [38]; (3) frequencies of non-ref NUMTs reported in Dayama G. et al. [39]; (4) frequencies of non-ref NUMTs detected in this study; and (5) arrows representing the insertion of mtDNA segments to the breakpoints of nDNA. e The number of somatic simple (left) and complex (right) SVs in 528 tumor samples with (num = 468) and without (num = 60) non-ref NUMTs. p values were calculated by the Wilcox test. f The circos plot of 46 distinct somatic non-ref NUMTs. g The circos plot of 22 tumor specific non-ref NUMTs. The genes with non-ref NUMTs on the intron region were labeled. Only the frequency in our cohort was shown in f and g. h (Left, middle) The expression of CBWD1 in different sample types in patients with different sources of the non-ref NUMTs chrM:6225–6420-to-chr9-129767. (Right) The expression of CBWD1 in samples with and without the non-ref NUMTs. p values were calculated by the Wilcox test