Nature Genetics: doi: /ng Supplementary Figure 1

Transcription

1 Supplementary Figure 1 LD (r 2 ) between the A3AB deletion and all markers in a 400-kb APOBEC3 region in 1000 Genomes Project populations. Populations: CEU, individuals of European ancestry from Utah, only samples that overlapped with the HapMap set are used here; CHB, Chinese from Beijing; JPT, Japanese from Tokyo. All samples were genotyped with a CNV assay; genotypes for all other markers were generated by the 1000 Genomes Project (October 2014 release). SNP rs is the only marker that tags the deletion in Europeans and Japanese (r 2 = 1.0) and Chinese (r 2 = 0.95). In the Yoruba (YRI) panel of the 1000 Genomes Project, the CNV is weakly polymorphic (4.2%) while rs is monomorphic, and LD metrics could not be calculated. SNP rs was also genotyped by a custom TaqMan assay (Supplementary Note), and all genotypes were 100% concordant with data from the 1000 Genomes Project.

2 Supplementary Figure 2 LD (r 2 ) between the A3AB deletion, its proxy SNP rs , and all GWAS-genotyped and imputed markers within a 400-kb APOBEC3 genomic region on chromosome 22q13.1. The plot is based on 848 genotyped or imputed markers in 1,837 samples from individuals of European ancestry from the PLCO study in which the deletion (gray box) was genotyped by a CNV assay and its proxy SNP, rs , was genotyped by a TaqMan genotyping assay. In this set, the CNV and rs have r 2 = 0.92 and D = Because of low LD with other markers in this region (best r 2 ~0.2), the deletion and its proxy SNP, rs , cannot be imputed and have to be genotyped.

3 Supplementary Figure 3 Electrophoretic mobility shift assays for SNP rs with nuclear extracts from bladder cancer cell line HTB-9 and breast cancer cell lines MDA-MB-231 and T-47D.

4 Supplementary Figure 4 Expression of selected APOBEC3 genes (A3A, A3B, and A3G) in GTEx. Expression analysis in 8,555 samples (53 normal human tissues from 544 donors) based on data generated by the Genotype-Tissue Expression (GTEx) Project. Expression is measured by RNA seq and presented as normalized log 10 (FPKM) values. Expression in bladder and breast tissue samples is marked by red boxes. Data for colon transverse tissue were available only for A3A and are labeled separately, while data for expression of A3A were not available for adipose visceral tissue.

5 Supplementary Figure 5 Expression of selected APOBEC3 genes (A3A, A3B, and A3G) in bladder cancer cell lines RT-4 and HTB-9 infected with Sendai virus (SeV) or treated with the DNA-damaging drug bleomycin (Bleo). (a,c) Increase in expression of a viral-specific RNA shows that cells were successfully infected with SeV. (b,d) In untreated cells, baseline expression of A3A is significantly lower than that of A3B and A3G. (e,g) A3A and A3G but not A3B are significantly induced after 12 h of SeV infection. (f,h) Expression of A3A, A3B, and A3G is significantly induced by 24 h of treatment with bleomycin as compared to untreated (UT) samples. Plots present expression values ( C t, log 2 scale) for targets (A3A, A3B, and A3G) normalized by the geometric mean of expression for two endogenous controls (GAPDH and PPIA). Dotted lines indicate the lower level of detection for the targets a C t value of 40 was assigned to samples for which expression was not detected by 40 cycles of qrt PCR; individual plot points for these samples are defined by the levels of expression of endogenous controls. All experiments were performed in biological triplicate. P values are for two-sided t tests. Shown are values for individual replicates and means. Raw data are available in Supplementary Data 2.

6 Supplementary Figure 6 Expression of selected APOBEC3 genes (A3A, A3B, and A3G) in breast cancer cell lines MDA-MB-231 and T-47D infected with Sendai virus or treated with the DNA-damaging drug bleomycin. (a,c) Increase in expression of a viral-specific RNA shows that cells were successfully infected with SeV. (b,d) In untreated cells, A3B expression is significantly higher than that of A3A and A3G. (e,g) Only A3A in MDA-MB-231 cells and all APOBEC genes in T-47D cells are significantly induced after 12 h of SeV infection. (f,h) Only A3B and A3G in MDA-MB-231 cells and all APOBEC genes in T-47D cells are significantly induced by 24 h of treatment with bleomycin as compared to untreated (UT) samples. Plots present expression values ( C t, log 2 scale) for targets (A3A, A3B, and A3G) normalized by the geometric mean of expression for two endogenous controls (GAPDH and PPIA). Dotted lines indicate the lower level of detection for the targets a C t value of 40 was assigned to samples for which expression was not detected by 40 cycles of qrt PCR; individual plot points for these samples are defined by the level of expression of endogenous controls. All experiments were performed in biological quadruplicate. P values are for two-sided t tests. Shown are values for individual replicates and means. NE, not expressed in all samples. Raw data are available in Supplementary Data 2.

7 Supplementary Figure 7 APOBEC mutagenesis and SNP rs as predictors of overall survival for patients with breast cancer in TCGA. (a h) Results are presented separately for patients with ER + (a d) and ER (e h) tumors. (a,e) Overall survival in relation to quartiles of APOBEC-signature mutation counts. (b,f) Overall survival in relation to APOBEC mutagenesis pattern classified as no or yes (at least one mutation present). (c,g) Overall survival in relation to APOBEC mutagenesis pattern classified as no, low mutation counts (1 48 mutations) or high mutation counts ( 49 mutations; based on the median in bladder tumors, presented in Fig. 5b). (d,h) Overall survival in relation to rs Hazards ratios and P values are for multivariate Cox regression models that also include age and tumor stage as core variables.

8 Supplementary Information Contents (Middlebrooks, Banday et. al.) Section detail Page Supplementary Tables: Supplementary Table 1 Supplementary Table 2 Supplementary Table 3 Supplementary Table 4 Supplementary Table 5 Supplementary Table 6 Supplementary Table 7 Supplementary Table 8 Supplementary Table 9 Supplementary Table 10 Supplementary Table 11 Association with bladder cancer risk for top genotyped or imputed markers within 1 Mb of the chr22q13.1 region (separate Excel file) Genotypes of A3AB deletion (CNV) and SNP rs in HapMap populations (separate Excel file) Association of SNP rs and A3AB deletion with bladder cancer risk in 1,719 bladder cancer cases and 2,566 controls of European ancestry from NCI-GWAS1 and 1,116 bladder cancer cases and 945 controls from Japan Exploratory analysis of association between SNP rs and expression of all gene isoforms within a 400 Kb APOBEC3 region in 357 bladder tumors in TCGA Exploratory analysis of association between SNP rs and expression of all gene isoforms within a 400 Kb APOBEC3 region in 541 breast tumors in TCGA Association of CpG site methylation with expression of major A3B isoform (uc003awo.1) in 357 bladder tumors in TCGA Association of all bladder cancer GWAS signals with counts of APOBEC-signature mutations in 357 bladder tumors in TCGA Effect of mrna expression of all APOBEC3 isoforms on APOBEC mutagenesis in TCGA bladder tumors Effect of mrna expression of all APOBEC3 isoforms on APOBEC mutagenesis in TCGA breast tumors Induction of known interferon-stimulated genes (ISGs) by infection with Sendai virus (SeV) or treatment with a DNAdamaging drug bleomycin (Bleo) in bladder cancer cell line HTB-9 and breast cancer cell line MCF7 APOBEC mutagenesis, SNP rs and expression of APOBEC3 isoforms as predictors of overall survival (OS) of bladder cancer patients in TCGA

9 Supplementary Table 12 Supplementary Table 13 Supplementary Table 14 Supplementary Table 15 Supplementary Table 16 Supplementary Table 17 Supplementary Table 18 Effect of somatic mutations in TP53 on overall survival (OS) of bladder cancer patients predicted by APOBEC mutagenesis and SNP rs in TCGA Distribution of somatic mutations in PIK3CA in TCGA bladder tumors in rs genotype groups Effect of somatic mutations in TP53 on overall survival (OS) of breast cancer patients predicted by APOBEC mutagenesis and SNP rs in TCGA Distribution of A3AB deletion genotypes in controls of European ancestry Distribution of A3AB deletion genotypes in controls from Japan Demographic, clinical, and genetic data for bladder and breast cancer patients in TCGA Correlation of A3AB deletion status with expression of main A3A and A3B isoforms in subsets of bladder and breast tumors in TCGA Supplementary Table 19 Distribution of A3AB deletion genotypes in subsets of bladder and breast tumors in TCGA and in HapMap samples Supplementary Table 20 Oligos for EMSA probes 23 Supplementary Note: Association of SNP rs with breast cancer risk in Breast Cancer Association Consortium (BCAC) 24 SNP rs as A3AB deletion proxy 25 Detailed analysis of survival in TCGA breast cancer patients 26 URLs 27 References 27 2

10 Supplementary Tables Supplementary Table 3. Association of SNP rs and A3AB deletion with bladder cancer risk in 1,719 bladder cancer cases and 2,566 controls of European ancestry from NCI-GWAS1 and 1,116 bladder cancer cases and 945 controls from Japan European ancestry* 1,719 cases and 2,566 controls Japanese 1,116 cases and 945 controls Variant Alleles, protective underlined Protective allele, % cases /controls OR (95%CI), P-value Adjusted for rs Adjusted for A3AB deletion Protective allele, % cases/ controls OR (95%CI), P-value Adjusted for rs Adjusted for A3AB deletion rs T/C 28.1/ ( ) P=3.13E ( ) P=1.02E / ( ) P=7.50E ( ) P=2.00E-02 A3AB deletion I/D 5.8/ ( ) P= ( ) P= / ( ) P= ( ) P= Results are for a dominant genetic model for A3AB deletion genotypes: deletion absence (I/I) vs. deletion presence (I/D or D/D) and for an additive genetic model for SNP rs *ORs and p-values are adjusted for age, sex, study site (SPBC, Spain and PLCO, USA) and smoking status (ever/never) 3

11 Supplementary Table 4. Exploratory analysis of association between SNP rs and expression of all gene isoforms within a 400 Kb APOBEC3 region in 357 bladder tumors in TCGA Gene Isoform ID, UCSC* Isoform annotation Beta-coefficient P-value APOBEC3A uc011aob.1 minor isoform APOBEC3A uc003awn.2 major isoform APOBEC3A uc011aoc.1 A3AB deletion isoform APOBEC3B uc003awo.1 major isoform E-05 APOBEC3B uc003awp.1 minor isoform APOBEC3B uc003awq.1 minor isoform APOBEC3C uc003awr APOBEC3D uc011aoe APOBEC3D uc011aod APOBEC3D uc003awt APOBEC3F uc003aww APOBEC3F uc011aog APOBEC3F uc003awv APOBEC3G uc003awy APOBEC3G uc003awx APOBEC3H uc011aoh APOBEC3H uc003axa APOBEC3H uc011aoi CBX6 uc003awm CBX6 uc003awl CBX7 uc003axc CBX7 uc003axb DNAL4 uc003awj NPTXR uc003awk SUN2 uc003awh SUN3 uc003awi SUN4 uc010gxq SUN6 uc011aoa *UCSC - University of California Santa Cruz genome browser. Beta-coefficients represent increase or decrease of quantile-normalized 1 log10 mrna expression per risk allele of SNP rs , adjusting for age, sex, and race. Bolded results are significant after Bonferroni multiple test correction (alpha = 1.79E-03). 4

12 Supplementary Table 5. Exploratory analysis of association between SNP rs and expression of all gene isoforms within a 400 Kb APOBEC3 region in 541 breast tumors in TCGA Gene Isoform ID, UCSC* Isoform annotation Beta-coefficient P-value APOBEC3A uc011aob.1 minor isoform APOBEC3A uc003awn.2 major isoform APOBEC3A uc011aoc.1 A3AB deletion isoform APOBEC3B uc003awo.1 major isoform E-03 APOBEC3B uc003awp.1 minor isoform APOBEC3B uc003awq.1 minor isoform APOBEC3C uc003awr APOBEC3D uc011aoe APOBEC3D uc011aod APOBEC3D uc003awt APOBEC3F uc003aww APOBEC3F uc011aog APOBEC3F uc003awv APOBEC3G uc003awy APOBEC3G uc003awx APOBEC3H uc011aoh APOBEC3H uc003axa APOBEC3H uc011aoi CBX6 uc003awm CBX6 uc003awl CBX7 uc003axc CBX7 uc003axb DNAL4 uc003awj NPTXR uc003awk SUN2 uc003awh SUN3 uc003awi SUN4 uc010gxq SUN6 uc011aoa *UCSC - University of California Santa Cruz genome browser. Beta-coefficients represent increase or decrease of quantile-normalized 1 log10 mrna expression per risk allele of SNP rs , adjusting for age and race. Bolded results are significant after Bonferroni multiple test correction (alpha = 1.79E- 03). 5

13 Supplementary Table 6. Association of CpG site methylation with expression of major A3B isoform (uc003awo.1) in 357 bladder tumors in TCGA CpG site Beta-coefficient P-value cg * E-08 cg E-07 cg E-05 cg E-05 cg E-04 cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg cg * Beta-coefficients represent increase or decrease of quantile-normalized 1 log10 mrna expression per log10 DNA methylation levels of CpG sites tested, not adjusting for any covariates. 6

14 Supplementary Table 7. Association of all bladder cancer GWAS signals with counts of APOBECsignature mutations in 357 bladder tumors in TCGA GWAS marker Region/ Gene Reference TCGA proxy SNP r 2 with GWAS marker Betacoefficients P- value rs q13.1/ APOBEC3 Rothman et al rs E-06 rs q12.3/ SLC14A2 Garcia-Closas et al rs rs q12.3 SLC14A2 Garcia-Closas et al rs rs p12.2 Figueroa et al rs rs p22.3/ CDKAL1 Figueroa et al rs rs p22/ Rothman et al NAT rs rs rs rs q28/ TP63 5p15.33/ TERT- CLPTML 11p15.5/ LSP1 Kiemeney et al rs Rafnar et al. rs Figueroa et al rs rs rs rs q12.3/ Rothman et al. CCNE q12.3/ Fu et al., CCNE q37.1/ Rothman et al. UGT1A rs rs rs rs p16.3/FGFR, TMEM129- TACC Kiemeney et al rs

15 rs q24.1 Kiemeney et al rs rs q26.2/ MYNN Figueroa et al rs rs q37.1/ Tang et al. UGT1A rs rs p12.2 Rafnar et al rs rs p12.2 Figueroa et al rs rs q34/ MCF2L Figueroa et al rs rs q24.3/ Wu et al. PSCA rs Beta-coefficients represent increase or decrease of log10 APOBEC-signature mutation counts per risk allele of candidate SNPs, adjusting for age, gender, and race. Bolded results are significant after Bonferroni multiple test correction (alpha = 2.50E-03). 8

16 Supplementary Table 8. Effect of mrna expression of all APOBEC3 isoforms on APOBEC mutagenesis in TCGA bladder tumors Dependent Variable: APOBEC mutagenesis pattern (log10_apobec_mutload_minestimate) Predictors Beta-coefficient P-value* Age Gender Race Tumor stage rs _genotypes (0, 1, 2) E-04 log_uc003awn2_a3a_major log_uc011aob1_a3a_minor log_uc011aoc1_a3ab_del log_uc003awo1_a3b_major E-04 log_uc003awp1_a3b_minor log_uc003awq1_a3b_minor log_uc003awr2_a3c log_uc003awt3_a3d log_uc011aod1_a3d log_uc011aoe1_a3d log_uc003awv2_a3f log_uc003aww2_a3f log_uc011aog1_a3f log_uc003awx2_a3g log_uc003awy2_a3g log_uc003axa3_a3h log_uc011aoh1_a3h log_uc011aoi1_a3h *P-values are for a multivariate linear regression model that includes all the variables listed; the results may differ from those presented in Figure 2 M-N which are based on a limited set of variables. Splicing forms are designated based on information from the UCSC genome browser. 9

17 Supplementary Table 9. Effect of mrna expression of all APOBEC3 isoforms on APOBEC mutagenesis in TCGA breast tumors Dependent Variable: APOBEC mutagenesis pattern (log10_apobec_mutload_minestimate) Predictors Beta-coefficient P-value* Age Race Tumor stage rs _genotypes (0, 1, 2) log_uc003awn2_a3a_major log_uc011aob1_a3a_minor log_uc011aoc1_a3ab_del log_uc003awo1_a3b_major log_uc003awp1_a3b_minor log_uc003awq1_a3b_minor log_uc003awr2_a3c log_uc003awt3_a3d log_uc011aod1_a3d log_uc011aoe1_a3d log_uc003awv2_a3f log_uc003aww2_a3f log_uc011aog1_a3f log_uc003awx2_a3g log_uc003awy2_a3g log_uc003axa3_a3h log_uc011aoh1_a3h log_uc011aoi1_a3h *P-values are for a multivariate linear regression model that includes all the variables listed; the results may differ from those presented in Figure 2 O-P which are based on a limited set of variables. Splicing forms are designated based on information from the UCSC genome browser. Expression of all isoforms is presented as quantile-normalized 1 log10 expression values. 10

18 Supplementary Table 10. Induction of known interferon-stimulated genes (ISGs) by infection with Sendai virus (SeV) or treatment with a DNA-damaging drug bleomycin (Bleo) in bladder cancer cell line HTB-9 and breast cancer cell line MCF7 SeV vs. Control, 12 hrs Bleo vs. Control, 24 hrs ISGs HTB-9 MCF-7 HTB-9 MCF-7 Gene Log2 p-val Log2 p-val Log2 p-val Log2 p-val MX E E E E-01 ISG E E E E-01 CXCL E E E-01 ND ND CCL E E E E-01 TLR E E E-04 ND ND IL E E E-04 ND ND STAT E E E E-02 IFNB E E E E-01 IL E E E E-01 MEFV E E-03 ND ND ND ND IFIH E E E E-02 DHX E E E E-01 OAS E E E E-01 AIM E E E-01 ND ND TNF E E E-03 ND ND CXCL E E E-01 ND ND NLRP E E E-04 ND ND CASP E E E-02 ND ND AZI E E E E-01 IL E E E E-01 APOBEC3G E E E-01 ND ND NFKBIA E E E E-01 IFNA E E E-01 ND ND CTSS E E E E-01 IL12A E E E-02 ND ND IL1B E E E-03 ND ND TICAM E E E E-02 MYD E E E E-01 CXCL E E E E-01 IRF E E E E-01 TRIM E E E E-02 CASP E E E E-01 CYLD E E E E-01 11

19 NFKB E E E E-01 MAP2K E E E E-01 IRF E E E E-01 CTSL E E E E-01 MAP2K E E E E-01 JUN E E E E-01 MAPK E E E E-02 MAPK E E E E-01 TRAF E E E E-01 CHUK E E E E-02 FOS E E E E-01 SPP E-01 ND ND E E-01 PYDC E E E E-02 Expression was quantified using antiviral qrt-pcr arrays (Qiagen), analysis is based on 2 biological replicates for each condition, p-values are for two-sided T-test. ND expression is not detected. For SeV experiment control group represents non-infected samples; for Bleo treatment control group represents treatment with DMSO (vehicle). Log2 values represent increase or decrease of expression (positive or negative values, respectively) compared to control groups. 12

20 Supplementary Table 11. APOBEC mutagenesis, SNP rs and expression of APOBEC3 isoforms as predictors of overall survival (OS) of bladder cancer patients in TCGA Treatment (yes/no) YES (N=81) NO (N=146) Adjustments* Predictor Betacoeff coeff Beta - P-value P-value Age, years Variables included Gender in all models Stage 1.02 <5.6E-03 rs Signature mutations E Mutagenesis pattern E E-03 log_uc003awn2_a3a_major log_uc011aob1_a3a_minor log_uc011aoc1_a3ab_del log_uc003awo1_a3b_major log_uc003awp1_a3b_minor log_uc003awq1_a3b_minor log_uc003awr2_a3c log_uc003awt3_a3d log_uc011aod1_a3d log_uc011aoe1_a3d log_uc003awv2_a3f log_uc003aww2_a3f log_uc011aog1_a3f log_uc003awx2_a3g log_uc003awy2_a3g log_uc003axa3_a3h log_uc011aoh1_a3h log_uc011aoi1_a3h Age, years Variables included Gender in all models Stage 0.77 <4.2E-04 rs Signature mutations E Mutagenesis pattern log_uc003awn2_a3a_major log_uc011aob1_a3a_minor log_uc011aoc1_a3ab_del log_uc003awo1_a3b_major log_uc003awp1_a3b_minor log_uc003awq1_a3b_minor log_uc003awr2_a3c log_uc003awt3_a3d log_uc011aod1_a3d log_uc011aoe1_a3d

21 ALL (N=356) With treatment info (N=227) log_uc003awv2_a3f log_uc003aww2_a3f log_uc011aog1_a3f log_uc003awx2_a3g log_uc003awy2_a3g log_uc003axa3_a3h log_uc011aoh1_a3h log_uc011aoi1_a3h Age, years Gender Stage <1E-06 Variables included in all models rs Signature mutations E E-04 Mutagenesis pattern E E-03 log_uc003awo1_a3b_major log_uc003awn2_a3a_major log_uc011aoc1_a3ab_del Age, years Variables included Gender in all models Stage 0.86 <2E-06 rs Signature mutations E E-03 Mutagenesis pattern E E-03 log_uc003awo1_a3b_major log_uc003awn2_a3a_major log_uc011aoc1_a3ab_del Treatment (Yes/No) Treatment: neoadjuvant, adjuvant, or radiation; ALL: all patients regardless of treatment information; With treatment info: only patients that have treatment information. All multivariate modes include age, gender and tumor stage as core variables. Adjustments include mutagenesis pattern for rs and rs for APOBEC mutagenesis variables. Negative beta-coefficients represent decreased hazard of death (longer survival) per bladder cancer risk allele rs a, with higher APOBEC mutagenesis metrics (log10 signature mutation counts or mutagenesis pattern), or higher expression of specific isoforms (quantile-normalized 1 log10 expression values), tested in separate models. Positive betacoefficients represent decreased survival with age, and increased tumor stage; treatment is not significantly associated with improved survival. Analyses are limited by samples with all the covariates available. Analysis in 73 patients with muscle-invasive bladder cancer treated with adjuvant platinumbased therapy showed significantly increased survival with increased expression of A3A, A3D and A3H in tumors 12. This is consistent with results presented above in TCGA bladder cancer patients with muscleinvasive bladder cancer who received treatment (predominantly adjuvant) survival is improved with increased APOBEC mutagenesis, expression of some isoforms of A3D, A3H and A3A (non-significant). The pattern is absent in TCGA bladder cancer patients who did not receive treatment. 14

22 Supplementary Table 12. Effect of somatic mutations in TP53 on overall survival (OS) of bladder cancer patients predicted by APOBEC mutagenesis and SNP rs in TCGA TP53 mutations SNP rs Combined Yes No A-cancer risk allele N (%) N (%) N (%) P-value* AA 139 (38.3) 83 (46.4) 56 (30.4) AG 152 (41.9) 67 (37.4) 85 (46.2) GG 72 (19.8) 29 (16.2) 43 (23.4) Total *Chi-Square test (df=2) for distribution of somatic mutations (Yes/No, not limited by APOBECsignature type) in TP53 in rs genotype groups. Predictor Beta-coefficient P-value rs E-02 rs * E-02 Signature mutations E-05 Signature mutations* E-05 Mutagenesis pattern E-05 Mutagenesis pattern* E-04 Cox regression models for overall survival (OS) are based on rs , log10 APOBEC mutagenesis metrics, adjusting for age, gender, tumor stage, with or without adjustment for presence of TP53 mutations*. Analysis is based on 356 samples with data available for all the variables. 15

23 Supplementary Table 13. Distribution of somatic mutations in PIK3CA in TCGA bladder tumors in rs genotype groups PIK3CA mutations SNP rs Combined Yes No A-cancer risk allele N (%) N (%) N (%) P-value AA 139 (38.3) 39 (47.6) 100 (35.6) 0.037* AG 152 (41.9) 34 (41.5) 118 (42.0) GG 72 (19.8) 9 (11.0) 63 (22.4) Total *Chi-Square test (df=2) for distribution of somatic mutations (Yes/No, not limited by APOBEC-signature type) in PIK3CA in rs genotype groups. Mutation data for TCGA samples were obtained from Firebrowse (Materials and Methods). We tested a panel of genes frequently mutated in bladder tumors (TP53, RB1, ELF3, TSC1, PIK3CA, RHOB, CDKN2A, ARID1A, ZFP36L1, CDKN1A, ATM and FGFR3) in relation to rs genotype groups. Significant enrichments of mutations in rs genotype groups were found for PIK3CA and TP53 (Supplementary Table 12). 16

24 Supplementary Table 14. Effect of somatic mutations in TP53 on overall survival (OS) of breast cancer patients predicted by APOBEC mutagenesis and SNP rs in TCGA ER+ breast tumors, n=387 TP53 mutations SNP rs Combined Yes No A-cancer risk allele N (%) N (%) N (%) P-value AA 123 (31.8) 19 (25.7) 104 (33.2) 0.158* AG 197 (50.9) 37 (50.0) 160 (51.1) GG 67 (17.3) 18 (24.3) 49 (15.7) Total Predictor Beta-coefficient P-value rs rs * Signature mutations Signature mutations* Mutagenesis pattern Mutagenesis pattern* ER- breast tumors, n=119 TP53 mutations SNP rs Combined Yes No P-value A-cancer risk allele N (%) N (%) N (%) AA 41 (34.5) 33 (41.3) 8 (20.5) 0.057* AG 61 (51.3) 36 (43.8) 26 (66.7) GG 17 (14.3) 12 (15.0) 5 (12.8) Total Predictor Beta-coefficient P-value rs rs * Signature mutations Signature mutations* Mutagenesis pattern Mutagenesis pattern* *Chi-Square test (df=2) for distribution of somatic mutations (Yes/No, not limited by APOBEC-signature type) in TP53 in rs genotype groups. Cox regression models for overall survival (OS) prediction based on rs , log10 APOBEC mutagenesis metrics, adjusting for age, and tumor stage, with or without adjustment for presence of TP53 mutations*. Analyses are limited by samples with data available for all the variables, n=387 for ER+ and n=119 for ER- tumors. 17

25 Supplementary Table 15. Distribution of A3AB deletion genotypes in controls of European ancestry A3AB deletion genotype PLCO N (%) SBCS N (%) Combined N (%) HapMap, Europeans* I/I 1,386 (85.87) 835 (87.71) 2,221 (86.55) 52 (86.67) I/D 220 (13.63) 116 (12.18) 336 (13.09) 8 (13.33) D/D 8 (0.50) 1 (0.11) 9 (0.35) 0 (0.00) Total 1, , PLCO - Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial, USA SBCS - Spanish Bladder Cancer Study A3AB deletion alleles: I - insertion, D deletion; * individuals of European ancestry from Utah (CEU); HapMap and study samples were genotyped for the deletion using the same CNV assay, HapMap genotypes were 100 % concordant with those reported earlier based on PCR and gel electrophoresis method

26 Supplementary Table 16. Distribution of A3AB deletion genotypes in controls from Japan A3AB deletion genotype Japanese N (%) HapMap, Japanese* I/I 495 (52.38) 20 (44.44) I/D 379 (40.11) 18 (40.00) D/D 71 (7.51) 7 (15.56) Total A3AB deletion alleles: I - insertion, D deletion; * Japanese from Tokyo, Japan (JPT); HapMap and study samples were genotyped for the deletion using the same CNV assay, HapMap genotypes were 93% concordant with those reported earlier based on PCR and gel electrophoresis method

27 Supplementary Table 17. Demographic, clinical, and genetic data for bladder and breast cancer patients in TCGA TCGA variables 357 bladder tumors N (%) 533 breast tumors N (%) Age (mean ± SD) ± ± Male 266 (74.5) 5 (0.94) Female 91 (25.5) 528 (99.1) Caucasian 294 (82.3) 445 (83.5) Asian 43 (12.0) 32 (6.0) African American 20 (5.6) 56 (10.5) Tumor stage T0 N/A N/A T1 2 (0.56) 84 (15.8) T2 116 (32.5) 302 (56.8) T3 120 (33.6) 137 (25.8) T4 114 (31.9) 5 (0.94) Unknown 5 (1.4) 4 (0.75) rs G/G 68 (19.0) 85 (15.9) A/G 151 (42.3) 269 (50.6) A/A G A 138 (38.7) 287 (40.2) 427 (59.8) 179 (33.5) 439 (41.2) 627 (58.8) The set is limited by samples with available data for age, sex, race, tumor stage, SNP rs , CpG site methylation and A3B CNA used for analyses presented in Figure 2 A-D. 20

28 Supplementary Table 18. Correlation of A3AB deletion status with expression of main A3A and A3B isoforms in subsets of bladder and breast tumors in TCGA Bladder, N=105 Gene Isoform ID, UCSC* Isoform annotation Beta-coefficient P-value A3A uc003awn.2 major isoform A3A uc011aoc.1 A3AB deletion isoform E-04 A3B uc003awo.1 major isoform Breast, N=388 Gene Isoform ID, UCSC* Isoform annotation Beta-coefficient P-value A3A uc003awn.2 major isoform E-03 A3A uc011aoc.1 A3AB deletion isoform E-05 A3B uc003awo.1 major isoform E-11 *UCSC - University of California Santa Cruz genome browser Analysis was performed in subsets of TCGA samples with available A3AB deletion genotype data 14. A3AB expression levels were tested for correlation with specified isoforms, adjusting for age, sex, and race. Deletion status, which is scored on DNA level as presence of 0, 1, or 2 deletion alleles significantly correlates with presence of A3AB deletion isoform annotated on mrna level, thus these variables can be used as proxies for each other. Incomplete correlation could be due to misclassification of A3AB deletion status or isoform annotation in TCGA. Expression of all isoforms is presented as quantile-normalized 1 log10 expression values. 21

29 Supplementary Table 19. Distribution of A3AB deletion genotypes in subsets of bladder and breast tumors in TCGA and in HapMap samples Bladder, N =105 Breast, N=388 Population I/I I/D and D/D Total MAF MAF in HapMap Caucasian (CEU) African American (YRI) Asian (CHB, JPT) Total Population I/I I/D D/D Total MAF MAF in HapMap Caucasian (CEU) African American (YRI) Asian (CHB, JPT) Total A3AB deletion alleles: I - insertion, D - deletion; HapMap populations: CEU individuals of European ancestry from Utah; YRI Yoruba from Nigeria, CHB Chinese from Beijing, JPT Japanese from Tokyo. HapMap samples were genotyped for A3AB deletion using a CNV assay (Table S2); breast and bladder cancer TCGA samples 14 were scored by mapping short exome-sequencing DNA reads to the reference genome. The deletion is more common in Asians than in other ethnic groups. 22

30 Supplementary Table 20. Primers SNP Oligo Sequence rs rs _g-f ATAAGGGCGTTGGGCAAGGAAA EMSA rs _g-r TTTCCTTGCCCAACGCCCTTAT rs _a-f ATAAGGGCGTTAGGCAAGGAAA rs _a-r TTTCCTTGCCTAACGCCCTTAT rs rs _a-f AGGTACTCCCAACCCCTGCAGC EMSA rs _a-r GCTGCAGGGGTTGGGAGTACCT rs _g-f AGGTACTCCCGACCCCTGCAGC rs _g-r GCTGCAGGGGTCGGGAGTACCT rs rs _g-f GAAGCGGCAGGGCCAGCCATGTG EMSA rs _g-r CACATGGCTGGCCCTGCCGCTTC rs _a-f GAAGCGGCAGGACCAGCCATGTG rs _a-r CACATGGCTGGTCCTGCCGCTTC qrt-pcr for SeV-DI RNA SeV-DI RNA_F SeV-DI RNA_R GTCAAGATGTTCGGGGCCAG CGTTCTGCACGATAGGGACT 23

31 Supplementary Note: Association of SNP rs with breast cancer risk in Breast Cancer Association Consortium (BCAC) 15. Estimates are based on 45,290 cases and 41,880 controls of European ancestry from 41 studies genotyped as part of the Collaborative Oncological Gene-environment Study (COGS) using a custom Illumina iselect genotyping array. 24

32 SNP rs TaqMan assay information This SNP is located in an intronic region upstream of exon 4 of A3A; the deletion starts after this exon. The region has high similarity to A3B and A3G, unique positions between these regions are marked in blue, position of TaqMan probe is in underlined italics. Assay specificity is defined by primers that uniquely recognize A3A and not A3B or A3G regions. rs a/c A3A: gatgtgggaagtctgtcctgagagtcatgggccctaggtgccaccccgatcccacagcgggagcgtgactta A3B: gatgtgggaagtctgtcctgagagtcatgggcccttggtgctgccccctccccacaacaggagcgtgactta A3G: gatgtgggaagtctgtcttgagagtcatgggccttggtgccaccacgat cccacagcgggagtgtgactta SNP rs TaqMan amplicon, 171 bp ATTCCAATGGGAAGGAACTGCCTGATGAAGGAGCTAAGTCCCTAGGGGAGGGAGAGGGAAAGGAGGGACTGAAACCAGGATG TGGGAAGTCTGTCCTGAGAGTC[A/C]TGGGCCCTAGGTGCCACCCCGATCCCACAGCGGGAGCGTGACTTATCTCCCCTGTCCCTT TTCAGA Forward primer: ATTCCAATGGGAAGGAACTGC Reverse primer: TCTGAAAAGGGACAGGGGAGA Probe_allele_A_VIC: GAGAGTCATGGGCCCTA Probe_allele_C_FAM: GAGAGTCCTGGGCCCTA Haplotypes for SNP rs and deletion (CNV) in PLCO set, n=1,837, D =0.97, r 2 =0.92 Haplotype rs deletion (CNV) Frequency A_I C_D A_D C_I Rs alleles A and C, A3AB deletion alleles: I insertion and D - deletion; concordance in duplicated samples: for rs % in 751 samples, for CNV 99.7% in 397 samples. Since CNV and rs are in strong linkage disequilibrium (LD) in our set of individuals of European ancestry (n = 1,837, D = 0.97, r 2 = 0.92), we used rs genotypes when CNV data could not be generated due to insufficient DNA quantity or quality (471 of 4,285 European samples, 11%). Detailed analysis of survival in TCGA breast cancer patients 25

33 ER+ tumors Predictors Beta-coefficient P-value Age, years E-05 Tumor stage E-04 log_nc_uc003awn2_a3a_major log_nc_uc011aob1_a3a_minor log_nc_uc011aoc1_a3ab_del log_nc_uc003awo1_a3b_major log_nc_uc003awp1_a3b_minor log_nc_uc003awq1_a3b_minor APOBEC mutagenesis pattern rs genotypes (0, 1, 2 risk alleles) Survival is in ER+ breast cancer is decreased with higher age, tumor stage and expression of A3AB isoform (A3AB germline deletion). Expression of all isoforms is presented as quantile-normalized 1 log10 expression values. ER- tumors Predictors Beta-coefficient P-value Age, years Tumor stage E-05 log_nc_uc003awn2_a3a_major log_nc_uc011aob1_a3a_minor log_nc_uc011aoc1_a3ab_del log_nc_uc003awo1_a3b_major log_nc_uc003awp1_a3b_minor log_nc_uc003awq1_a3b_minor APOBEC mutagenesis pattern rs genotypes (0, 1, 2 risk alleles) Survival is in ER- breast cancer is decreased with higher tumor stage and expression of minor form of A3B. Expression of all isoforms is presented as quantile-normalized 1 log10 expression values. 26

34 URLs: GTEx: Breast Cancer Association Consortium data: apps.ccge.medschl.cam.ac.uk/consortia/bcac/) REFERENCES: 1. Shabalin, A.A. Matrix eqtl: ultra fast eqtl analysis via large matrix operations. Bioinformatics 28, (2012). 2. Rothman, N. et al. A multi-stage genome-wide association study of bladder cancer identifies multiple susceptibility loci. Nat Genet 42, (2010). 3. Garcia-Closas, M. et al. A genome-wide association study of bladder cancer identifies a new susceptibility locus within SLC14A1, a urea transporter gene on chromosome 18q12.3. Hum Mol Genet 20, (2011). 4. Figueroa, J.D. et al. Genome-wide association study identifies multiple loci associated with bladder cancer risk. Hum Mol Genet 23, (2014). 5. Kiemeney, L.A. et al. Sequence variant on 8q24 confers susceptibility to urinary bladder cancer. Nat Genet 40, (2008). 6. Rafnar, T. et al. Sequence variants at the TERT-CLPTM1L locus associate with many cancer types. Nat Genet 41, (2009). 7. Fu, Y.P. et al. The 19q12 bladder cancer GWAS signal: association with cyclin E function and aggressive disease. Cancer Res 74, (2014). 8. Kiemeney, L.A. et al. A sequence variant at 4p16.3 confers susceptibility to urinary bladder cancer. Nat Genet 42, (2010). 9. Tang, W. et al. Mapping of the UGT1A locus identifies an uncommon coding variant that affects mrna expression and protects from bladder cancer. Hum Mol Genet 21, (2012). 10. Rafnar, T. et al. Genome-wide association study yields variants at 20p12.2 that associate with urinary bladder cancer. Hum Mol Genet 23, (2014). 11. Wu, X. et al. Genetic variation in the prostate stem cell antigen gene PSCA confers susceptibility to urinary bladder cancer. Nat Genet 41, (2009). 12. Mullane, S.A. et al. Correlation of Apobec Mrna Expression with overall Survival and pd-l1 Expression in Urothelial Carcinoma. Sci Rep 6, (2016). 13. Kidd, J.M., Newman, T.L., Tuzun, E., Kaul, R. & Eichler, E.E. Population stratification of a common APOBEC gene deletion polymorphism. PLoS Genet 3, e63 (2007). 14. Nik-Zainal, S. et al. Association of a germline copy number polymorphism of APOBEC3A and APOBEC3B with burden of putative APOBEC-dependent mutations in breast cancer. Nat Genet 46, (2014). 15. Michailidou, K. et al. Large-scale genotyping identifies 41 new loci associated with breast cancer risk. Nat Genet 45, , 361e1-2 (2013). 27