Improved statistical methods for identifying generic risk factors underlying complex diseases

糖尿病、哮喘和癌症等复杂疾病可能涉及多个基因，这些基因与生活方式选择和环境暴露共同作用，增加个人患疾病的风险。通过研究家族遗传来寻找潜在基因的努力一直受到低功率的阻碍，因此只取得了有限的成功。因此，生物医学研究人员现在正转向人群研究，将疾病或其他与疾病相关的特征的存在与DNA序列的变异联系起来。收集这样的关联数据是昂贵的。如何才能最大限度地利用这些数据资源？对数据进行有效的统计分析是关键。例如，在确定一个人的疾病风险时，研究人员如何以最佳方式提取关于基因是否与环境相互作用的信息？另一个问题是如何处理基因组数据，即所谓的单倍型，由于所有个体都是远亲，这些数据不完全被观察到，并具有复杂的潜在依赖关系。最后，我们如何才能有效地将以前从动物或细胞等其他类型的研究中获得的信息纳入对人类群体的研究？这项研究将调查所有这三个统计问题，并开发和传播分析工具，作为免费提供的软件，供更广泛的生物医学界使用。在这项研究过程中开发的统计方法和软件将使用正在进行的生物医学合作者研究的数据进行验证和改进。在国际和加拿大国内，在遗传流行病学和相关研究领域，如统计学和人口遗传学方面，严重缺乏能力。因此，加拿大卫生研究所的遗传学研究所宣布遗传流行病学是国家培训的战略重点。拟议的研究将提供这一应用战略领域的统计基础方面的培训机会。拥有这种统计基础的高素质人员将处于有利地位，为造福下一代加拿大人的跨学科健康研究做出贡献。

Complex disorders such as diabetes, asthma and cancer likely involve a number of genes acting in concert with lifestyle choices and environmental exposures to increase an individual's risk to develop disease. Efforts to find the underlying genes by studying inheritance in families have been hampered by low power and, consequently, have met with only limited success. As a result, biomedical investigators are now turning to population studies that correlate presence of the disease or other disease-related traits to variation in DNA sequences. Collecting such association data is expensive. How can we make the best use of these data resources? Efficient statistical analyses of the data are key. For example, how can investigators optimally extract information about whether or not genes interact with the environment in determining a person's disease risk? Another question is how to handle genomic data, known as haplotypes, that are incompletely observed and have complex underlying dependencies due to the fact that all individuals are distantly related. Finally, how can we effectively incorporate into studies of human populations previously available information from other types of studies such as those on animals or cells? This research will investigate all three of these statistical questions and develop and disseminate, as freely available software, analytical tools accessible to the wider biomedical community. Statistical methods and software developed during the course of this research will be validated and refined using data from ongoing studies of biomedical collaborators. Internationally and within Canada, there is a critical lack of capacity in genetic epidemiology and related research areas such as statistical and population genetics. As a result, the Institute of Genetics at the Canadian Institutes for Health Research has declared genetic epidemiology a national strategic priority for training. The proposed research will provide training opportunities in the statistical underpinnings of this applied strategic area. Highly qualified personnel with such a statistical foundation  will be well-placed to contribute to interdisciplinary health research benefitting the next generation of Canadians.