Synthetic mutational spectra with mixed and correlated mutational signatures SBS1 and SBS5

Experience suggests that it harder to extract mutational signatures that always co-occur and that generate correlated numbers of mutations. Here we provide 12 data sets, each consisting of 500 synthetic mutational spectra with varying degrees of mixture and correlation between two mutational signatures. The signatures studied were SBS1, a "clock-like" signature due to deamination of 5-methyl cytosine that consists primarily of mutations from CG > TG, and SBS5, a relatively flat "clock-like" signature. By "clock-like" we mean that the numbers of mutations attributable to these signatures increase with patients' ages. The 12 data sets varied in two dimensions: (i) average ratio of the number of SBS1 to the number of SBS5 mutations in spectra in the data set and (ii) correlation between the number of SBS1 and the number of SBS5 mutations in spectra in the data set. The mutational signatures SBS1 and SBS5 are described at https://doi.org/10.7303/syn12025148 and https://doi.org/10.1101/322859.