人类群体遗传结构的图论主成分分析方法

The Methodology of Graph Theory Principle Components Analysis on Human Population Genetic Structure

目的提出基因频率矩阵的图论主成分分析方法,探讨该法在人类群体遗传结构研究中的应用。方法从分析基因频率矩阵的结构特征入手,将主成分分析与图论中的最小生成树有机结合,构建人类群体遗传结构的图论主成分模型,并以中国26个汉族人群HLA-A基因座遗传结构分析为例,验证图论主成分分析的科学性和适用性。结果图论主成分分析的基本步骤可概括为:①对中心化基因频率的协方差矩阵进行主成分分析;②按图论原理求过m维空间n个点的最小生成树;③利用求“颈”法分割最小生成树;④将最小生成树整合到二维主成分散点图中构建图论主成分分类图。根据此步骤,对中国26个汉族人群HLA-A基因座遗传结构进行了图论主成分分析,分析结果符合中华民族源与流的客观规律。结论图论主成分分类图既可显示各群体的遗传结构特性,又可利用最小生成树的链接关系揭示各群体间的内在联系;图论主成分分析是分析人类群体遗传结构的一种较好方法。

Objective This paper presents the methodology of graph theory principle components analysis, and explores its applicability for studying human population genetic structure. Methods Based on the structure of gene frequency matrix, we combine the method of principle components analysis with the minimal spanning tree of graph theory, to set up the model of graph theory principle components analysis. As an example, the genetic structure of HLA-A locus in 26 Hun populations is analyzed by using graph theory principle components analysis, to show its rationality for studying human population genetic structure. Results The step by step of graph theory principle components analysis can be summarized as follows： ① Carry out principle components analysis to the centred gene frequency covariance matrix; ②Calculate the minimal spanning tree, which passes the points in dimensions space; ③Decompose the minimal spanning tree into several parts using the method of necks found; ④Set up the graph theory principal components classification graph in the principal components classification scallergram. According to these steps, the genetic structure of HLA- A locus in 26 Han populations is analyzed, which indicates that the results accord with the population genetic mechanism of Chinese Han population. Conclusion The graph theory principal components classification scallergram can show not only the genetic structure of populations, but also the intrinsic relationship among the populations. The graph theory principle components analysis is ideality method for studying the human population genetic structure.