抑郁症患者单核苷酸多态性(SNPs)分布特征的潜在类别分析

Latent Class analysis of SNPs Distribution Characters in Depression Patients

目的介绍潜在类别模型的原理及技术,应用此技术分析抑郁性疾病的单核苷酸多态性位点SNPs的潜在分布,探讨潜在类别间的差异与含义。方法采用Mplus软件,对抑郁患者单核苷酸多态性7个SNPs检测数据进行潜在类别分析。结果通过潜在类别分析把7个SNPs检测数据分为两个类别,类别1以杂合子为主,类别2以纯合子为主,结合个体特质应对特征发现,类别1具有消极应对高倾向性,而类别2具有消极应对低倾向性。结论潜在类别模型综合了结构方程模型与对数线性模型的思想,形成了自身的优势,其目的在于以最少的潜在类别数目来解释显变量之间的关联,由此提示我们潜在类别模型可以推广应用于基因组学与基因治疗等新兴领域。

Objective Introduce the principle and technique of latent class model（LCM）,use the technique to analysize latent distribution of SNPs,Explore the latent differences and the meaning between the latent class.Methods Apply Mplus software to analyze 7 SNPs by LCM.Results According to LCM,7 SNPs is divided into two clusters,cluster-1 mostly for heterozygous and cluster-2 for homozygous.Combined the personal trait,cluster-1 have a high-tendency to negative response,while the cluster-1 had a low-tendency to negative response inclination.Conclusion With the characteristics of structural equation model（SEM）and log-linear model,LCM has formed its own advantage and its aim is to explain the significant correlation between variables by the least number of latent clusters,suggests that LCM can be promoted to be used in genomics and gene therapy and other new area.