Pooled Analyses of SNPs and Risk of Head & Neck Cancer

描述(申请人提供)：以前的头颈癌分子流行病学研究已经在数百例患者和对照中检测了一到几个单核苷酸多态(SNPs)，重点是致癌物代谢和DNA修复基因的序列变异。在这些研究中观察到的不一致结果可能是由于：i)在检测适度风险序列变体方面的统计能力较低，ii)假阳性结果，iii)发表偏倚，以及iv)每个SNP单独赋予显著风险增加的适度先验概率。此外，考虑到致癌是高度复杂的，许多基因作用于多个途径，加上它们与环境因素的相互作用，检查一个或几个适度风险序列变体的主要影响是一种过于简化的方法。从单标记分析转向多个SNPs和基因-环境相互作用的测试是必要的，但需要相当大的样本量和统计能力。因此，我们建议在国际头颈癌流行病学(INHANCE)联盟内进行汇集分析，该联盟是最近成立的一个研究小组合作，领导头颈部癌症的大分子流行病学研究。我们将结合来自7项美国研究和3项欧洲研究的数据，这些研究涉及18个SNP的5000多个病例对照对。我们的具体目标是：1)对编码致癌物代谢酶和头颈癌风险的基因中的11个SNPs进行汇集分析；2)对DNA修复基因中的7个SNP进行汇集分析；3)应用新的统计方法，包括评估假阳性报告概率和分层建模方法，以解决假阳性结果的可能性并纳入关于这些基因的先验知识。令人感兴趣的假设是：a)致癌物新陈代谢和DNA修复基因中的SNPs会增加患头颈癌的风险，并且 每个SNP的影响不大，b)这些基因座上的多个等位基因的遗传赋予了更高的头颈癌风险，c)这些序列变异体和环境因素之间存在着多种基因-基因和基因-环境的相互作用。我们相信，我们建议的分析是为大型头颈癌研究生成数据的一种经济有效的方法，这将解决先前研究的局限性，并提供关于这些SNP在头颈癌发生中的作用的更清晰的图景。

DESCRIPTION (provided by applicant): Previous molecular epidemiology studies on head and neck cancer have examined one to several single nucleotide polymorphisms (SNPs) among several hundred cases and controls, focusing on sequence variants in carcinogen metabolism and DNA repair genes. The inconsistent results observed across these studies may be due to: i) low statistical power in detecting modest risk sequence variants, ii) false positive results, iii) publication bias, and iv) a moderate prior probability that each SNP individually confers substantial increase in risk. Furthermore, examining the main effects of one or few modest risk sequence variants is an overly simplified approach, when considering that carcinogenesis is highly complex, with numerous genes acting on multiple pathways plus their interactions with environmental factors. Moving from single marker analyses to tests of multiple SNPs and gene-environment interactions is necessary, but requires considerable sample size and statistical power. Therefore, we propose to conduct pooled analyses within the International Head and Neck Cancer Epidemiology (INHANCE) Consortium, a recently established collaboration of research groups leading large molecular epidemiology studies of head and neck cancer. We will combine data from 7 US studies and 3 European studies on more than 5000 case-control pairs for 18 SNPs. Our specific aims are: 1) to conduct pooled analysis on 11 SNPs in genes that encode carcinogen metabolizing enzymes and the risk of head and neck cancer, 2) to conduct pooled analysis on 7 SNPs in DNA repair genes, 3) to apply novel statistical methods including the assessment of false positive report probability and hierarchical modeling approaches to address the possibility of false positive results and to incorporate prior knowledge on these genes. The hypotheses of interest are: a) SNPs in the carcinogen metabolism and DNA repair genes confer an increased risk of head and neck cancer, and the magnitude of effect of each SNP is modest b) inheritance of multiple alleles at these loci confer a higher risk of head and neck cancer, c) there are multiple gene-gene and gene-environment interactions among these sequence variants and environmental factors. We believe that our proposed analyses is a cost effective approach in generating data for a large head and neck cancer study, that will address the limitation of previous studies and provide a clearer picture of the role of these SNPs in head and neck cancer carcinogenesis.