Cross Cancer Institute, Alberta Health Services, Edmonton, Alberta, Canada.
Eur J Hum Genet. 2012 Jun;20(6):682-9. doi: 10.1038/ejhg.2011.273. Epub 2012 Jan 18.
Genome-wide association studies for breast cancer have identified over 40 single-nucleotide polymorphisms (SNPs), a subset of which remains statistically significant after genome-wide correction. Improved strategies for mining of genome-wide association data have been suggested to address heritable component of genetic risk in breast cancer. In this study, we attempted a two-stage association design using markers from a genome-wide study (stage 1, Affymetrix Human SNP 6.0 array, cases=302, controls=321). We restricted our analysis to DNA repair/modifications/metabolism pathway related gene polymorphisms for their obvious role in carcinogenesis in general and for their known protein-protein interactions vis-à-vis, potential epistatic effects. We selected 22 SNPs based on linkage disequilibrium patterns and high statistical significance. Genotyping assays in an independent replication study of 1178 cases and 1314 controls were attempted using Sequenom iPLEX Gold platform (stage 2). Six SNPs (rs8094493, rs4041245, rs7614, rs13250873, rs1556459 and rs2297381) showed consistent and statistically significant associations with breast cancer risk in both stages, with allelic odds ratios (and P-values) of 0.85 (0.0021), 0.86 (0.0026), 0.86 (0.0041), 1.17 (0.0043), 1.20 (0.0103) and 1.13 (0.0154), respectively, in combined analysis (N=3115). Of these, three polymorphisms were located in methyl-CpG-binding domain protein 2 gene regions and were in strong linkage disequilibrium. The remaining three SNPs were in proximity to RAD21 homolog (S. pombe), O-6-methylguanine-DNA methyltransferase and RNA polymerase II-associated protein 1. The identified markers may be relevant to breast cancer susceptibility in populations if these findings are confirmed in independent cohorts.
全基因组关联研究已确定了超过 40 个单核苷酸多态性（SNP），其中一部分在全基因组校正后仍具有统计学意义。为了解决乳腺癌遗传风险的可遗传性部分，已经提出了改进的全基因组关联数据挖掘策略。在这项研究中，我们使用全基因组研究中的标记（Affymetrix Human SNP 6.0 阵列，病例=302，对照=321）进行了两阶段关联设计。我们将分析仅限于与 DNA 修复/修饰/代谢途径相关的基因多态性，因为它们在一般致癌作用中具有明显作用，并且与已知的蛋白质-蛋白质相互作用有关，具有潜在的上位效应。我们根据连锁不平衡模式和高统计学意义选择了 22 个 SNP。在使用 Sequenom iPLEX Gold 平台进行的 1178 例病例和 1314 例对照的独立复制研究中尝试进行了基因分型检测（第 2 阶段）。在两个阶段中，6 个 SNP（rs8094493、rs4041245、rs7614、rs13250873、rs1556459 和 rs2297381）均显示出与乳腺癌风险一致且具有统计学意义的关联，其等位基因比值（和 P 值）分别为 0.85（0.0021）、0.86（0.0026）、0.86（0.0041）、1.17（0.0043）、1.20（0.0103）和 1.13（0.0154），在合并分析（N=3115）中。其中，三个多态性位于甲基-CpG 结合域蛋白 2 基因区域，并且紧密连锁。其余三个 SNP 靠近 RAD21 同源物（酿酒酵母）、O-6-甲基鸟嘌呤-DNA 甲基转移酶和 RNA 聚合酶 II 相关蛋白 1。如果这些发现在独立队列中得到证实，那么所鉴定的标记可能与人群中的乳腺癌易感性有关。
