Misawa Kazuharu
Department of Human Genetics Yokohama City University Graduate School of Medicine 3-9 Fukuura, Kanazawa-ku Yokohama 236-0004 Japan.
Adv Genet (Hoboken). 2022 Apr 5;3(3):2100066. doi: 10.1002/ggn2.202100066. eCollection 2022 Sep.
Recent advances in sequencing technologies enable genome-wide analyses for thousands of individuals. The sequential kernel association test (SKAT) is a widely used method to test for associations between a phenotype and a set of rare variants. As the sample size of human genetics studies increases, the computational time required to calculate a kernel is becoming more and more problematic. In this study, a new method to obtain kernel statistics without calculating a kernel matrix is proposed. A simple method for the computation of two kernel statistics, namely, a kernel statistic based on a genetic relationship matrix (GRM) and one based on an identity by state (IBS) matrix, are proposed. By using this method, calculation of the kernel statistics can be conducted using vector calculation without matrix calculation. The proposed method enables one to conduct SKAT for large samples of human genetics.
测序技术的最新进展使得能够对数千人进行全基因组分析。序列核关联检验（SKAT）是一种广泛用于检验表型与一组罕见变异之间关联的方法。随着人类遗传学研究样本量的增加，计算核所需的计算时间变得越来越成问题。在本研究中，提出了一种无需计算核矩阵即可获得核统计量的新方法。提出了一种计算两种核统计量的简单方法，即基于遗传关系矩阵（GRM）的核统计量和基于状态相同（IBS）矩阵的核统计量。通过使用这种方法，核统计量的计算可以使用向量计算而无需矩阵计算。所提出的方法使人们能够对大量人类遗传学样本进行SKAT。
