Statistical, population genetics and genetic epidemiology

一个人的基因构成可以被认为塑造了他们患复杂疾病的倾向，而环境因素引发了疾病的发生，并与遗传因素一起，可以改变疾病的发展。我的团队的研究反映了我们继续参与大规模遗传和基因组研究的设计和分析。我们继续设计不仅对遗传应用有用的方法，而且普遍适用于正在评估许多统计假设的其他类型的多维数据的分析，例如，它适用于暴露的表观遗传效应研究、代谢组学研究和差异基因表达研究。我们的方法被设计为通用的，因为它们计算P值，这些值可能来自各种测试，包括那些考虑环境暴露和基因环境交互作用的测试。我们一直在扩展我们研究的合作方面，努力将现代统计方法引入到与复杂疾病和环境暴露有关的研究中。在Slade等人(2013)中，我们进行了一项病例对照遗传关联研究，研究了2925个SNPs与一种常见的慢性面部疼痛状况--颞下颌关节紊乱病(TMD)的两种亚型之间的关系：1)局限性TMD；2)TMD伴广泛疼痛。使用我们的统计扩展到通路分析，我们发现，与健康对照组相比，局限性TMD患者映射到5-羟色胺受体途径的SNPs的等位基因频率不同，而具有广泛疼痛的TMD病例映射到T细胞受体途径的SNPs的等位基因频率不同。代表来自5-羟色胺能途径的6个SNP的综合效应的风险指数与局部性TMD的几率更大相关，该结果在重复病例对照队列研究中重现。代表来自T细胞受体途径的8个SNPs的综合效应的风险指数与TMD和广泛性疼痛的几率更大相关。5-羟色胺能信号与局限性TMD之间的重复关联代表了关于慢性疼痛病因的新发现，而T细胞受体途径与TMD与广泛性疼痛之间的关联证实了最近关于T细胞介导的机制在疼痛中的作用的研究。在人类相关性研究中，常见的遗传变异通常有助于分子病理生理过程，在人类相关性研究中，SNPs与临床疾病之间的关联很弱。在这项研究中，我们区分了慢性TMD的生物相关亚型的表型，并应用了一种新的分析方法，寻找已知细胞信号通路上通常较小的遗传影响的组合。我们发现，解剖学上的局部性和全身性疼痛代表了TMD的临床意义亚群，它们具有不同的分子特征和相应的不同遗传背景。这些结果揭示了5-羟色胺能通路在TMD的病理生理学中的明显作用，尽管还需要进一步的工作来揭示这一贡献的方向。5-羟色胺受体2选择性配体可能有效地治疗这一导致局限性TMD发生的主要细胞途径。 TMD项目促使我们开始研究设计方法，在这种设计中，研究的第二部分(例如，复制)可以利用初步研究中观察到的效果方向(Kuo等人，2013年)。初步研究提供了关于每个SNP上的哪个等位基因增加疾病易感性的信息(即影响方向)。因此，对于复制样本中的特定个体，可以为每个相关的SNP添加易感等位基因的数量，从而产生单个分数。基于这样的分数的统计将独立于在主要研究中使用的关联统计，即使它借用了主要研究的效果方向，只要在主要研究中使用的检验是双侧的(即，忽略效果方向)。因此，可以添加主要和次要(例如，复制)研究的关联统计(具有反映研究大小的权重)，并且可以确定加权和的分布。我们之前的研究表明，仔细汇集不同样本的汇总统计数据，基本上可以与基于个人层面数据的分析一样强大。此外，在某些情况下，基于两种统计数据(来自主要研究和次要研究)的组合测试可能比基于所有数据的单一测试更强大。这其中的一些原因包括纳入了从主要研究中借用效果方向所产生的附加信息，以及第二研究中的分数测试是单自由度测试的事实。如果第一次研究中观察到的所有效果方向都是正确的，那么基于单一分数的第二次研究测试将具有非常高的统计能力。然而，由于统计上的不确定性，一些影响方向可能是不正确的。正确测定的数量的可变性也受到连锁不平衡的影响。因此，我们调查了这些不确定性会在多大程度上影响电力。我们还研究了关于权力的初级和次级研究对象的最优分配问题。

The genetic makeup of a person can be thought of as shaping their propensity to complex diseases, while environmental factors trigger onset of diseases and, together with genetic factors, can modify their progression. Research of my group reflects our continuing involvement in the design and analysis of large-scale genetic and genomic studies. We continue to devise methodology that is useful not only for genetic applications but also generally applicable for analysis of other kinds of multidimensional data where many statistical hypotheses are being evaluated, e.g. it is applicable in studies of epigenetic effects of exposures, in metabolomics studies, and in studies of differential gene expression. Our methods are designed to be general because they work off P‑values, which can result from various tests, including those accounting for environmental exposures and gene‑environment interactions. We have been extending collaborative aspects of our research, striving to bring modern statistical methodology to studies concerned with complex diseases and environmental exposures. In Slade et al (2013) we conducted a case-control genetic association study of relationships between 2925 SNPs and two subtypes of a commonly occurring chronic facial pain condition, temporomandibular disorder (TMD): 1) localized TMD; and 2) TMD with widespread pain. Using our statistical extensions to pathway analysis we found that when compared to healthy controls, cases with localized TMD differed in allelic frequency of SNPs that mapped to a serotonergic receptor pathway, while cases of TMD with widespread pain differed in allelic frequency of SNPs that mapped to a T-cell receptor pathway. A risk index representing combined effects of six SNPs from the serotonergic pathway was associated with greater odds of localized TMD, and the result was reproduced in a replication case-control cohort study. A risk index representing combined effects of eight SNPs from the T-cell receptor pathway was associated with greater odds of TMD with widespread pain. The replicated association between serotonergic signaling and localized TMD represents a novel finding regarding etiology of chronic pain, while the observed association between T cell receptor pathway and TMD with widespread pain corroborates recent studies of the role of T cell-mediated mechanisms in pain. Common genetic variants usually contribute to molecular pathophysiological processes in ways that produce weak associations between individuals SNPs and clinical disease when assessed in human association studies. In this research, we distinguished phenotypically between biologically relevant subtypes of chronic TMD and applied a new analytic approach that searches for combinations of usually small genetic influences on known cellular signaling pathways. We found that anatomically localized and generalized pain represent clinically meaningful subgroups of TMD and that they have distinct molecular profiles with correspondingly-distinct genetic backgrounds. The results reveal a distinct role for serotonergic pathways in pathophysiology of TMD, although further work is needed to uncover the direction of this contribution. Potentially, this major cellular pathway contributing to development of localized TMD might be effectively treated using serotonin receptor 2 selective ligands. The TMD project prompted us to start looking into methodology for designs where the second (e.g. replication) part of the study can take advantage of effect directions observed in the primary study (Kuo et al., 2013). The primary study gives information about which allele at each SNP increases susceptibility to disease (i.e. the effect direction). Thus, for a particular individual in the replication sample, the number of susceptibility alleles can be added for each associated SNP, resulting in a single score. A statistic based on such a score would be independent of an association statistic used in the primary study, even though it borrows effect directions of the primary study, as long as the test utilized in the primary study is two-sided (i.e. oblivious to the effect direction). Therefore, association statistics of the primary and the secondary (e.g. replication) studies can be added (with weights reflecting the study sizes), and the distribution of the weighted sum can be determined. Our previous research suggests that careful pooling of summary statistics from different samples can be essentially as powerful as analysis based on individual-level data. Moreover, it is possible that in some situations, the test based on a combination the two statistics (from the primary and the secondary study) can be more powerful than a single test based on all data. Some reasons for this include incorporation of additional information that results from borrowing effect directions from the primary study and the fact that the score test in the secondary study is a single degree of freedom test. If all effect directions observed in the primary study were always correct, the secondary study test based on a single score would have had a remarkably high statistical power. However, due to statistical uncertainty, some effect directions may be incorrect. The variability in the number of correct determinations is also affected by linkage disequilibrium. Therefore, we investigated to what extent the power would be affected by these uncertainties. We also studied the problem of optimal allocation of subjects to the the primary and the secondary study with regard to power.