Fig. 1

Definition and statistics of SNPs. a For SNPs in each sample, we require the human reference sequence to be identical to the sequence in macaque (Macaca mulatta) so that the alternative allele would be the derived allele. This is the rational way to confirm the direction of mutations. b To define the tumor-specific mutations, we require the normal samples to have only the ancestral allele and the tumor sample to have both ancestral and derived alleles. Sites with insufficient sequencing coverages in normal samples were uncertain and excluded (grey). Sites with 100% mutation levels in tumor samples were also unreliable and discarded (green). Non-ancestral sites (purple, very few) were not considered. c Fraction of SNPs regarding their functional annotation. d Percentage of SNPs regarding the nucleotide change of the mutation