人类遗传流行病学的研究设计:海外朝鲜民族研究方案（英文）

传统的遗传流行病学和基因定位的研究设计往往局限于单一的（通常是简单的）数据结构和过分简化的流行病学模型。然而，当实际情况稍微复杂一些的时候，这些简单的方法几乎注定要失败，因此，我们应该重新建构思维方式，应考虑何种的病因学模型会与我们所知道的进化，表现型的表达以及基因型-表现型的相互关系相符，当基因定位的策略充分利用选择偏倚来使P（基因型Ⅰ表现型的；选择偏倚）的预测值达到最大，用分离分析试图对外显率函数或P（表现型Ⅰ基因型）作无偏估计，因而一般而言，没有一个值可以完全充分地被另一个值所推估，一个常见的，关于基因组计划得以“改变世界”的误解甚至使得严谨的科学家们都忽略了以下的科学事实；他们努力地为基因定位策略上的巨额支出辩护，却忽略了基因定位策略的本身就是极度取决于一些难以验证的前题假设，将不同实验设计方案的数据合并起来定会使估计参数和检验假设的自由度增加，虽然仍然需要大样本数的资料库，但采取以上策略可提供更多的信息以检测出真实情况，以往传统的研究方法只针对病因变异量的某一特定方面；双生子法侧重于研究在一个家庭内分享共同的环境因素之下遗传因素的作用；寄养子研究侧重于在控制遗传因素之下，不同家庭环境间的变异性；移民流行病学分析侧重于在控制遗传因素之下，不同文化环境间的变异性；而家系分析则侧重于检测较大的亲缘族系中的复杂遗传性相关，不同的研究着重不同成份的变异量，很显然，一个综合性的研究将可对特微病因学提供更多关于自然，养育因素以及它们之间交互作用的信息，同样地，这些信息可用来辅助基因定位和基因型-表现型相关性的研究，我们目前正以外朝鲜族家庭为对象进行此类研究，这些外朝鲜族家庭包括位于哈萨克斯坦，韩国，中国（延边0的朝鲜族家庭以及寄养于美国及欧洲的朝鲜人，我们亦希望能取得这些寄养人位于韩国的血亲的资料。

Study Design for Genetic Epidemiology and Gene Mapping:The Korean Diaspora Project

Study designs in genetic epidemiology and gene mapping have traditionally focused on single (often simple) data structures with oversimplified epidemiological models. However, when the true state of nature is even slightly more complex, these simple methods will fail almost surely. To this end, it may be important to restructure our thinking in terms of what sorts of etiological models would be consistent with what we know about evolution, secular trends in phenotypic expression, and the ephemeral nature of the genotype phenotype relationships, among other things. Segregation analysis attempts to make unbiased estimates of the penetrance functions, or P(Phenotype| Genotype) while gene mapping strategies make full use of ascertainment bias to maximize the predictive value, P(Genotype| Phenotype; Ascertainment), such that neither can be predicted meaningfully from the other in general1]. Popular misconceptions about the world changing effects" of the genome project have lead even serious scientists to ignore inconvenient scientific facts in an effort to justify expenditures of large sums of money on mapping strategies that are critically dependent on difficult to justify assumptions. Combining data from different study designs will certainly increase the number of degrees of freedom available for estimation of parameters and testing of hypotheses in a given dataset.Naturally the size of the dataset also needs to be large, but such approaches are potentially much more informative about the true state of nature. Each traditional study design focuses on one specific aspect of etiological variance.Twin studies focus on relative effects of genes conditional on shared environmental factors within a household.Adoption studies focus on inter household variation in environmental exposure whilepartially controlling for genetics.Migrant studies focus on intercultural variation in environmental exposures while controlling for population level genetic differences, and family studies focus on the complex sets of genetic correlations that can be measured within larger kindreds.The components of variance emphasized by each study are radically different, and it is clear that a study combining elements of each would certainly be more informative about the relative contributions of nature, nurture, and interactions between them on trait etiology. Similarly, this information can be used to assist in mapping strategies, and genotype phenotype correlation studies.We are currently developing such a study using populations of the Korean diaspora, composed of families of Koreans in Kazakstan, Korea, China (延边) and the USA, as well as internationally adopted Koreans in the USA and Europe, hopefully with data from their biological relatives in Korea as well.