Liu Yan, Foulkes Andrea S
Division of Biostatistics, 404 Arnold House, 715 North Pleasant Street, Amherst, MA 01003, USA.
Biom J. 2011 Sep;53(5):838-54. doi: 10.1002/bimj.201000218.
Characterizing associations among multiple single-nucleotide polymorphisms (SNPs) within and across genes, and measures of disease progression or disease status will potentially offer new insight into disease etiology and disease progression. However, this presents a significant analytic challenge due to the existence of multiple potentially informative genetic loci, as well as environmental and demographic factors, and the generally uncharacterized and complex relationships among them. Latent variable modeling approaches offer a natural framework for analysis of data arising from these population-based genetic association investigations of complex diseases as they are well-suited to uncover simultaneous effects of multiple markers. In this manuscript we describe application and performance of two such latent variable methods, namely structural equation models (SEMs) and mixed effects models (MEMs), and highlight their theoretical overlap. The relative advantages of each paradigm are investigated through simulation studies and, finally, an application to data arising from a study of anti-retroviral-associated dyslipidemia in HIV-infected individuals is provided for illustration.
表征基因内部和基因之间多个单核苷酸多态性（SNP）之间的关联，以及疾病进展或疾病状态的指标，可能会为疾病病因和疾病进展提供新的见解。然而，由于存在多个潜在的信息丰富的基因位点，以及环境和人口统计学因素，以及它们之间通常未被表征的复杂关系，这带来了重大的分析挑战。潜在变量建模方法为分析这些基于人群的复杂疾病基因关联研究产生的数据提供了一个自然的框架，因为它们非常适合揭示多个标记的同时效应。在本手稿中，我们描述了两种这样的潜在变量方法，即结构方程模型（SEM）和混合效应模型（MEM）的应用和性能，并强调了它们的理论重叠。通过模拟研究探讨了每种范式的相对优势，最后，提供了一个应用于HIV感染个体抗逆转录病毒相关血脂异常研究数据的示例。
