Hsieh Ai-Ru, Chen Dao-Peng, Chattopadhyay Amrita Sengupta, Li Ying-Ju, Chang Chien-Ching, Fann Cathy S J
Graduate Institute of Biostatistics, China Medical University, Taichung, Taiwan.
Institute of Biomedical Sciences, Academia Sinica, Nankang, Taipei, Taiwan.
PLoS One. 2017 Nov 30;12(11):e0188566. doi: 10.1371/journal.pone.0188566. eCollection 2017.
A region-specific method, NTR (non-threshold rare) variant detection method, was developed-it does not use the threshold for defining rare variants and accounts for directions of effects. NTR also considers linkage disequilibrium within the region and accommodates common and rare variants simultaneously. NTR weighs variants according to minor allele frequency and odds ratio to combine the effects of common and rare variants on disease occurrence into a single score and provides a test statistic to assess the significance of the score. In the simulations, under different effect sizes, the power of NTR increased as the effect size increased, and the type I error of our method was controlled well. Moreover, NTR was compared with several other existing methods, including the combined multivariate and collapsing method (CMC), weighted sum statistic method (WSS), sequence kernel association test (SKAT), and its modification, SKAT-O. NTR yields comparable or better power in simulations, especially when the effects of linkage disequilibrium between variants were at least moderate. In an analysis of diabetic nephropathy data, NTR detected more confirmed disease-related genes than the other aforementioned methods. NTR can thus be used as a complementary tool to help in dissecting the etiology of complex diseases.
开发了一种区域特异性方法，即NTR（非阈值罕见）变异检测方法——它不使用定义罕见变异的阈值，并考虑效应方向。NTR还考虑区域内的连锁不平衡，并同时兼顾常见和罕见变异。NTR根据次要等位基因频率和优势比权衡变异，将常见和罕见变异对疾病发生的影响合并为一个单一分数，并提供一个检验统计量来评估该分数的显著性。在模拟中，在不同效应大小下，NTR的功效随着效应大小的增加而增加，且我们方法的I型错误得到了很好的控制。此外，将NTR与其他几种现有方法进行了比较，包括组合多变量和压缩方法（CMC）、加权和统计方法（WSS）、序列核关联检验（SKAT）及其改进方法SKAT-O。在模拟中，NTR产生了相当或更好的功效，特别是当变异之间的连锁不平衡效应至少为中等程度时。在对糖尿病肾病数据的分析中，NTR检测到的已确认疾病相关基因比上述其他方法更多。因此，NTR可作为一种补充工具，有助于剖析复杂疾病的病因。
