Additional file 1: of Reduced intrinsic DNA curvature leads to increased mutation rate

Figure S1. Test of the Poisson distribution of nonsense mutations in URA3. Figure S2. The effect of nucleotide at the potential nonsense site on mutation rate. Figure S3. The estimation of the average mutation rate and the values of DNA properties. Figure S4. The results in cancer cells were not affected by SNV calling methods. Figure S5. DNA curvature is negatively associated with mutation rate in coding sequences in human tumors. Figure S6. The results in cancer cells held after controlling for the trinucleotide context. Figure S7. The results in cancer cells held after controlling for the heptanucleotide context. Figure S8. Logistic regression for predicting the presence of a SNV in human tumors. Figure S9. De novo point mutations in the human germline are enriched in DNA regions with a smaller DNA curvature. Figure S10. Expression levels were not significantly different among URA3 variants. Figure S11. Electrophoretic mobility shift assay showing differences in DNA curvature among URA3 variants. Figure S12. Comparison of intrinsic DNA curvature among yeast genes. Figure S13. Comparison of intrinsic DNA curvature among human genes. Figure S14. The distribution of the number of colonies on 60 5-FOA plates for each of the five URA3 variants. Table S1. Numbers of plates containing a mutation in URA3. Table S2. Models on predicting the mutation rate of a potential nonsense site in CAN1. Table S3. Modeling the mutation rate of a potential nonsense site in four human genes. Table S4. DNA sequences of URA3 variants. Table S5. Features of five URA3 variants. Table S6. Primers used in this study. (DOCX 1081 kb