Developing new statisical methods to detect variants involved in complex disease

鼻咽癌(NPC)具有独特的全球分布模式--东南亚和东半球的一些其他局部地区，这表明该疾病的风险是由环境暴露和特定遗传因素共同驱动的。早期的研究已经在基因组的不同区域发现了许多基因座。然而，这些关联的含义仍然是个谜。根据我们的初步结果，我们假设各种DNA修复基因的遗传变异和突变与鼻咽癌有关。此外，我们开发了更好的信息学工具来将DNA修复基因作为一个簇进行分析。 我们的结果可能为鼻咽癌的遗传决定因素提供新的见解，并可能为鼻咽癌和其他与感染相关的疾病的风险基因提供见解。我们从这项研究中学到的东西也可以应用于其他复杂疾病的基因研究，包括精神分裂症、躁郁症和自闭症等精神障碍。 DNA修复在防止环境致癌方面起着至关重要的作用，DNA修复基因的遗传变异已被报道与几种人类恶性肿瘤有关。为了评估DNA修复基因变异是否与鼻咽癌风险相关，我们在广东省内的广东人中国中进行了一项基于途径的候选基因关联研究。这项研究分两个阶段进行。在发现阶段，我们在配对的病例对照研究中调查了676个SNP(由包括姚博士在内的专家小组选择)，涵盖88个DNA修复基因，其中包括755例鼻咽癌患者和755例社区健康对照。我们发现RAD51L1、BRCA2和TP53BP1三个基因与鼻咽癌的发生发展密切相关。在验证阶段，我们跟踪了阶段1中揭示的信号，发现只有RAD51L1基因中的两个遗传变异与显著相关。验证的样本量为1,558例病例和1,297例对照。此外，我们还进行了基因-环境互作分析。我们第一次有证据表明，吸烟和食用咸鱼与DNA修复基因变异之间存在显著的相互作用。 在这项研究中，我们利用最先进的、高通量的基因分型方法以及复杂的遗传分析方法，除了单SNP分析外，还考虑了基因-基因相互作用和基因-环境相互作用。例如，我们进行了基于创新路径的分析，如CART。CART方法在危险因素和Tp53BP1、RAD51、BRCA2和MNA1基因的遗传变异中确定了鼻咽癌风险较高的亚组。这种相互作用通过我们自己的方法得到了证实。这一发现表明，吸烟和四个基因的变异可能共同导致鼻咽癌发病风险的增加。 在六个月内，我们成功地识别了有助于识别鼻咽癌高危人群的遗传标记。我们未来的工作将集中于在单独的人群中进一步验证这一发现，并通过结合第一阶段和第二阶段分析中使用的临床样本来识别有效预测鼻咽癌结果的遗传标记。基因和临床信息的结合可能比单独使用任何一种信息更具预测性，可能有利于使用适当的方法治疗鼻咽癌。

Nasopharyngeal carcinoma (NPC) has a unique global distribution pattern -- Southeast Asia and some other localized regions of the eastern hemisphere that suggests the disease risk is driven by a combination of environmental exposures and specific genetic factors. Earlier studies have implicated many loci in various regions of the genome. Nevertheless, the implications of these associations remain enigmatic. Based on our preliminary results, we hypothesized that genetic variants and mutations in various DNA repair genes are associated with NPC. Further, we developed better informatics tools for analyzing DNA repair genes as a cluster. The results we obtained may provide new insights into the genetic determinants of NPC, and may possibly provide insights on genes contributing to risk for NPC and other diseases related to infections. What we learn from this study can also be applied to genetic research for other complex diseases including psychiatric disorders including schizophrenia, bipolar and autism. DNA repair plays a critical role in protecting against environmental carcinogenesis, and genetic variants of DNA repair genes have been reported to be associated with several human malignancies. To assess whether or not DNA repair gene variants are associated with nasopharyngeal carcinoma (NPC) risk, we conducted a pathway based candidate gene association study among the Cantonese population within the Guangdong Province, China. The study was conducted in two stages. In the discover stage, we investigated 676 SNPs (selected by an expert panel including Dr. Yao) which cover 88 DNA repair genes in a matched case-controls study in which 755 patients who were diagnosed of NPC and 755 community based health controls were included. We discovered that three genes, RAD51L1, BRCA2 and TP53BP1 were significantly involved in development of NPC. In the validation stage, we followed up the signals revealed in stage 1 and found that only two genetic variants in the RAD51L1 gene were significantly associated. The sample size of the validation was 1,558 cases and 1,297 controls. Furthermore, we conducted gene-environment interaction analysis. For the first time, we documented the exposure of cigarette smoking and salted fish consumption had significant interactions with the DNA repair gene variations with confidence. In this study, we utilize state-of-the-art, high-throughput genotyping methods as well as sophisticated methods of genetic analysis that take into account gene-gene interaction and gene-environment interactions in addition to single-SNP analyses. For example, we conducted innovative pathway based analysis such as CART. The CART method identified a subgroup with higher risk of NPC among the risk factors and the genetic variants of the Tp53BP1, RAD51, BRCA2 and MNAT1 genes. This interaction was confirmed by our own method. This finding indicated that smoking and variants in four genes may contribute jointly to the increased risk of NPC development. Within six months, we successfully identified genetic markers that help identify those at risk for NPC development. Our future work will focus on further validating this finding in a separate population and identifying genetic markers that effectively predict NPC outcome by combing the clinical samples used in stage I and stage II analysis. The combination of genetic and clinical information may be more predictive than either alone, potentially benefiting the treatment of NPC using appropriate methods.