EBioMedicine (Oct 2019)

Re-evaluating genetic variants identified in candidate gene studies of breast cancer risk using data from nearly 280,000 women of Asian and European ancestryResearch in context

  • Yaohua Yang,
  • Xiang Shu,
  • Xiao-ou Shu,
  • Manjeet K. Bolla,
  • Sun-Seog Kweon,
  • Qiuyin Cai,
  • Kyriaki Michailidou,
  • Qin Wang,
  • Joe Dennis,
  • Boyoung Park,
  • Keitaro Matsuo,
  • Ava Kwong,
  • Sue Kyung Park,
  • Anna H. Wu,
  • Soo Hwang Teo,
  • Motoki Iwasaki,
  • Ji-Yeob Choi,
  • Jingmei Li,
  • Mikael Hartman,
  • Chen-Yang Shen,
  • Kenneth Muir,
  • Artitaya Lophatananon,
  • Bingshan Li,
  • Wanqing Wen,
  • Yu-Tang Gao,
  • Yong-Bing Xiang,
  • Kristan J. Aronson,
  • John J. Spinell,
  • Manuela Gago-Dominguez,
  • Esther M. John,
  • Allison W. Kurian,
  • Jenny Chang-Claude,
  • Shou-Tung Chen,
  • Thilo Dörk,
  • D. Gareth R. Evans,
  • Marjanka K. Schmidt,
  • Min-Ho Shin,
  • Graham G. Giles,
  • Roger L. Milne,
  • Jacques Simard,
  • Michiaki Kubo,
  • Peter Kraft,
  • Daehee Kang,
  • Douglas F. Easton,
  • Wei Zheng,
  • Jirong Long

Journal volume & issue
Vol. 48
pp. 203 – 211

Abstract

Read online

Background: We previously conducted a systematic field synopsis of 1059 breast cancer candidate gene studies and investigated 279 genetic variants, 51 of which showed associations. The major limitation of this work was the small sample size, even pooling data from all 1059 studies. Thereafter, genome-wide association studies (GWAS) have accumulated data for hundreds of thousands of subjects. It's necessary to re-evaluate these variants in large GWAS datasets. Methods: Of these 279 variants, data were obtained for 228 from GWAS conducted within the Asian Breast Cancer Consortium (24,206 cases and 24,775 controls) and the Breast Cancer Association Consortium (122,977 cases and 105,974 controls of European ancestry). Meta-analyses were conducted to combine the results from these two datasets. Findings: Of those 228 variants, an association was observed for 12 variants in 10 genes at a Bonferroni-corrected threshold of P < 2·19 × 10−4. The associations for four variants reached P < 5 × 10−8 and have been reported by previous GWAS, including rs6435074 and rs6723097 (CASP8), rs17879961 (CHEK2) and rs2853669 (TERT). The remaining eight variants were rs676387 (HSD17B1), rs762551 (CYP1A2), rs1045485 (CASP8), rs9340799 (ESR1), rs7931342 (CHR11), rs1050450 (GPX1), rs13010627 (CASP10) and rs9344 (CCND1). Further investigating these 10 genes identified associations for two additional variants at P < 5 × 10−8, including rs4793090 (near HSD17B1), and rs9210 (near CYP1A2), which have not been identified by previous GWAS. Interpretation: Though most candidate gene variants were not associated with breast cancer risk, we found 14 variants showing an association. Our findings warrant further functional investigation of these variants. Fund: National Institutes of Health. Keywords: Re-evaluation, Genetic variants, Candidate gene studies, Breast cancer risk