dc.description.abstract | To date, genome-wide association studies (GWAS) have identified common genetic variants at approximately 200 loci for breast cancer risk. GWAS have been predominately conducted among women of European ancestry, and the identified risk variants cannot be directly replicated among women of African ancestry primarily due to differences in genetic architectures. In this study, we performed a meta-analysis of GWAS for breast cancer in 18,044 cases and 22,187 controls of African ancestry. We identified common variants at 12 loci associated with risks of breast cancer overall and estrogen receptor (ER) status at the genome-wide significance level (P <5x10-8). Of them, two (rs10853615 at 18q11.2 and rs76664032 at 2q14.2) are novel with the sentinel variants located at least 1Mb away from any previously reported risk variants. Of the sentinel variants at known risk loci, four (located at 4q24, 6q25.1, 14q13.3, and 18q12.1) are not in linkage disequilibrium (r2 <0.02) in African-ancestry populations with the previously reported index variants. We developed a polygenetic risk score (PRS) using 109 variants selected in 17,279 cases and 21,422 controls, and validated it in an independent testing set (765 incident cases and 765 matched controls). In the testing set, the OR per standard deviation of the PRS was 1.54 (95% confidence interval [CI]: 1.38, 1.71), with an area under the receiver operating characteristic curve of 0.620 (95% CI: 0.591, 0.648). To identify putative susceptibility genes for breast cancer, we conducted a transcriptome-wide association study (TWAS). We built expression prediction models for 6,507 genes with a R2 of >0.01 and evaluated them in association analyses. Although none of these genes showed an association that reached Bonferroni-corrected significance level, ten previously reported genes showed a nominal significant association with breast cancer risk in the same direction as reported in previous TWAS in women of European ancestry. Our study identified novel risk variants for breast cancer, which provided additional insight into the genetic mechanisms of breast cancer etiology and improved risk prediction for this cancer among African-ancestry women. | |