稀有变异的有效发现与识别

Genome-wide association studies have discovered many genetic variants associated with human complex diseases. However, many of those variants can only explain a small portion of the human inherited risk. Recently, it is believed that rare variants account for some of the missing heritability for human complex diseases. Rare variants discovery and calling are mainly based on genome sequencing techniques. Since rare individuals account for only a small fraction of population, therefore thousands of individuals are required for a study to detect rare variants. Costs will be prohibitively high at this scale. In this project, based on the framework of pooling design, we will explore optimization design matrix which could efficiently improve rare variants discovery and calling and reduce cost as much as it can. Furthermore, we will make modification to improve our previous proposed variant identification by pooling algorithm (VIP) in terms of variant locus identification, variant allele frequency, and variant sample decoding.

基因组关联研究已经揭示了许多与人类复杂疾病相关联的遗传变异。但是人们已经认识到这些遗传变异仅能够解释人类复杂疾病遗传风险的很小的一部分。当前，稀有变异已经成为全面解释人类复杂疾病病因的重要组成部分，稀有变异的发现和识别主要基于基因组测序技术。由于稀有变异个体数占群体的比例非常小，因此需要通过对大量个体进行测序才能发现和识别稀有变异个体，这使得测序花费变得非常昂贵甚至不可行。本项目基于池设计方法，将通过定义最优的设计阵，从而能够最有效地发现和识别稀有变异，同时使花费变得最少。在此基础上，本项目将对我们之前提出的一种变异识别算法(Variant Identification by Pooling，VIP)做全面改进研究，具体地讲，对VIP在稀有变异位置的发现，稀有变异基因频率估计以及稀有变异个体的识别这三方面做改进研究。