Genetic Association Analysis of Lipid Profiles Using Linear Mixed Model.
10.4070/kcj.2006.36.3.229
- Author:
Kijun SONG
1
;
Chan Mi PARK
;
Kil Seob LIM
;
Yang Soo JANG
;
Dong Kee KIM
Author Information
1. Department of Biostatistics, Yonsei University College of Medicine, Seoul, Korea. dkkimbios@yumc.yonsei.ac.kr
- Publication Type:Original Article
- Keywords:
SNPs;
Lipids;
Association
- MeSH:
Apolipoproteins B;
Cardiovascular Diseases;
Confounding Factors (Epidemiology);
Genotype;
Humans;
Male;
Polymorphism, Single Nucleotide
- From:Korean Circulation Journal
2006;36(3):229-235
- CountryRepublic of Korea
- Language:Korean
-
Abstract:
BACKGROUND AND OBJECTIVES: Analyzing the association between multiple SNPs and the disease outcomes will provide new insight into the disease's etiology. However, this presents an analytic difficulty due to the large number of SNPs and the complex relationships among them. We proposed using the mixed model approach to identify the significant multi-locus genotypes and the high-order gene-to-gene interactions. SUBJECTS AND METHODS: We described the mixed effects model and applied this approach to real world data. For the purpose of these analyses, we examine the association of four types of SNPs (AGT5, APOB, CETP3 and ACE6) with the lipid profiles and the measures related with cardiovascular disease. We used data from 672 healthy individuals (283 males and 389 females) who were without cardiovascular diseases. RESULTS: The results of our analysis suggested that there were significant random genotype patterns and genotype groups according to the gender effect on the lipid profiles. In other words, there was significant variability across the genotype groups because of the effect of gender on the lipid profiles. CONCLUSION: The mixed model approach provided a flexible statistical framework for controlling potential confounding variables and for identifying a significant genetic contributions that may come about through the effects of multi-locus genotypes or through an interaction between the genotype and environmental variables (e.g. gender) with the variations in quantitative traits (e.g. lipid profiles). There were significant genetic contributions to the variability in the lipid profiles, and these were explained by the 4 SNPs described in our real data.
