Development of models in statistical genetics that incorporate spatio-temporal information

为了提高统计遗传学方法的能力，我们开发了利用时空结构信息的统计模型。通过开发的dn/ds比的空间聚集的伊辛模型之前，我们的分子进化模型检测到流感HA的受体结合区域上的多样化选择。基于观察到疟疾的种群内遗传多样性与撒哈拉以南非洲的地理距离之间的负相关性，我们通过考虑一维垫脚石来模拟疟疾在非洲以外的扩张，并估计殖民化率和当地的承载能力。通过开发一种圆形扫描统计方法来测量病例和对照之间罕见CNV所覆盖的基因数量的差异，我们从整个基因通路中提取了疾病相关的基因簇。

To improve the power of methods in statistical genetics, we developed statistical models that utilize the information of spatio-temporal structures. By developing the Ising model prior of the spatial aggregation of the dn/ds ratio, our model of molecular evolution detected the diversifying selection acting on the receptor-binding region of influenza HA. Based on the observation of the negative correlation between within-population genetic diversity of malaria and geographic distance from Sub-Saharan Africa, we modelled the expansion of malaria out of Africa by considering a one-dimensional stepping stone and estimated the rate of colonization and the local carrying capacities. By developing a circular scan statistics that measure difference in the number of genes intersected by rare CNVs between the cases and controls, we extracted disease-associated gene clusters from within a whole gene pathway.