| The genetics mechanics of complex trait include many genes. Multi-loci cis-acting sequences and trans-acting factors can interact to affect the complex trait. Therefore, compare to single allele analysis, multi-loci analysis will gain more information than individual SNP analysis. Haplotypes, the linear arrangement of alleles on the same chromosome inherited as a unit, provide a natural framework for testing the association between genetic markers and complex traits more efficiently than separate marker analysis. Obviously, haplotype analysis provides a mechanism for the use of SNP to research complex genetic traits more convenient and more efficient. A genetic architecture underlying the complex trait may exhibits strong multi-allele and multi-gene interaction effects within/among loci in presence of negligible main effects. Therefore, recent studies have not only demonstrated potential gains in statistical power by considering locus-by-locus interaction, but also detected gene-gene interactions contributing to disease-associated traits.Research contents are as followings:First, association between one-region SNPs and complex trait based on semiparametric regression modelAssociation mapping of complex traits typically employs SNP genotype data to identify a trait locus within a region of interest. However, considerable debate exists regarding the most powerful strategy for utilizing such SNP data for inference. A popular approach tests SNPs within the region individually, but such test would lose power as a result of incomplete linkage disequilibrium between the genotyped SNP. Alternatively, one can jointly test all SNPs simultaneously within the region by using haplotype, but if there are many haplotype, such multivariate tests have large freedom degrees that can also compromise power. In order to consider multiple SNPs simultaneously, but produce a test statistic with reduced degrees of freedom, basing on the shortcoming of kwee et al, we consider the genotype missing and get the best loci combination through step-down multiple comparison tests. Prostate cancer is common for human and threatens people's life. In this section, we use the public HapMap data which include LCLs of 67 genes and genotype of 339 candidate genes from CEU, CHB, JPT and YRI, to study the cis- and trans- regulatory in human prostate cancer pathway, and analyze pathway of the significance genes. Second, association between multi-region haplotype configuration and complex trait on parametric regressionIt is quite common that the genetic architecture of complex traits involves many genes and their interactions. Therefore, dealing with multiple unlinked genomic regions simultaneously is desirable. In this paper we develop a regression-based approach to assess the interactions of haplotypes that belong to different unlinked regions, and we use score statistics to test the null hypothesis of no genetic association. Additionally, multiple marker combinations at each unlinked region are considered. The multiple tests are settled via the minP approach. The P value of the"best"multi-region multi-marker configuration is corrected via Monte-Carlo simulations. Through simulation studies, we assess the performance of the proposed approach and demonstrate its validity and power in testing for haplotype interaction association. Our simulations showed that, for binary trait without covariates, our proposed methods prove to be equal and even more powerful than htr and hapcc which had been put in the FAMHAP program. Additionally, our model can be applied to a wider variety of traits and allows adjustment for other covariates. To test the validity, our methods are applied to analyze the association between four unlinked candidate genes and pig meat quality.Third, association between complex trait and multi-loci haplotype based on semiparametric regression modelThe genetic mechanism of complex traits includes many genes and their interaction. We bring forward a new method which considers a semiparametric model for complex trait association that uses haplotype genetic information from multiple genes simultaneously in analysis,but compare to existing multivariate approaches, produces a test statistic with reduced degrees of freedom. According to Liu et al and Kwee et al, we use LSKM to estimate the parameters and use score test to test the nonparameter function. In order to get the best genes interaction, we import step-down P value. Through simulation studies, we assess the performance of the proposed approach and demonstrate its validity and power in testing for genome wide association. We put this method application to analyze the association between KLK3 expressions and 339 candidate genes, and find the group of genes which affect these genes'expressions. The result shows that this analysis can obtain more information than above single gene analyzing. In addition, we apply this method to study the genetic mechanism of porcine meat.Forth, association between multi-region haplotype and binary trait with semiparametric logistic kernel machine regressionWith the advent of increasing efficient means to obtain genetic information, a great insurgence of data has resulted, which leads to the need for methods for analyzing the association of disease and the pathway. Because traditional parametric statistical approaches will produce high degrees, therefore, it is preferred to the nonparametric method. We propose a logistic kernel machine regression model for testing the association between binary outcomes and genome pathway. According to Liu et al, we turn our semiparametric model into logistic mixed model and use exiting statistic software to estimate the parameter, and use score test to test the nonparametric function. We assess the performance of the proposed approach and demonstrate its validity and power in testing for genome wide association. We apply to bisphosphonate-related osteonecrosis of the jaw in 2 serious of homogeneously treated MM patients.Fifth, association between multi-region haplotype and longitudinal trait by semiparametric regressionIt is very important to model the time and covariate effects on an outcome variable for studying longitudinal traits with several records. In view of repeated record for complex trait, we should consider the time effect for testing the association between the traits and haplotypes of genes. Based on the semiparametric model to research on the longitudinal trait of Zhang et al, we treat the haplotype and other covariates as fix effect. The parametric was estimated according to Zhang et al. In addition, we use likelihood ratio to testing the haplotypes effect. Through comparing this semiparametric model and common linear mixed model to analyze the haplotype, it is more appropriate to consider the repeated traits than single dot. We use this semiparametric model to analyze the association between the porcine reproduction which has repeated record and the haplotypes of two genes MMP1 and MMP10.In summary, in view of solving existing problem in genomics research, we set up generalized linear model to analysis muti-region haplotype, semiparametric model based on kernel function and son on. Through simulation studies, we assess the performance of the proposed approach and demonstrate its validity and power. We put application to several real data in reality and prove the simulation effect. These researches not only make progress on candidate gene of complex trait, but also provide a theoretical foundation for the implementation of quantitative trait, disease reseach and so on. In addition, we have put these methods on software and can be free download, which will give more chance for the researchers to carry out genome association.
|
PDF Full Text Request