Localizing Putative Markers in Genetic Association Studies by Incorporating Linkage Disequilibrium into Bayesian Hierarchical Models

Abstract

Objective: Numerous methods have been proposed to model the association between multiple single nucleotide polymorphisms (SNPs) and a phenotype. Often these methods do not explicitly model the information regarding the linkage disequilibrium (LD) between SNPs. Furthermore, many methods shrink the SNP effects towards zero, rather than to an unknown latent gene-level effect. Methods: We outline the use of bayesian hierarchical models for gene-level analysis that incorporates LD information. Four different bayesian models, with either the inclusion or exclusion of LD information and ‘shrinkage’ of SNP effects to a latent gene-level effect or zero, were applied to a pharmacogenomic study and simulation study. Results: We observed that the inclusion of LD information resulted in increased precision in the SNP parameter estimates. The simulation study also demonstrated that the bayesian models were able to determine the simulated ‘causative’ variant more often with less false-positive associations as compared to commonly used multi-SNP analysis methods. Conclusions: Incorporating LD information in the analysis of multiple SNPs results in more precise estimates of SNP effects. In addition, the bayesian models are able to isolate the simulated ‘causative’ variant more often than commonly used regression modeling methods.

A novel linkage-disequilibrium corrected genomic relationship matrix for SNP-heritability estimation and genomic prediction

“…This fact is true despite the strong assumption of equal weighting of loci in the G-BLUP model. Additionally, there are various WGR models proposed (Conti and Witte 2003 ; Sillanpää and Bhattacharjee 2005 ; Malo et al 2008 ; Tsai et al 2008 ; Fridley and Jenkins 2010 ; Yang and Tempelman 2012 ; Yi et al 2015 ) in order to account for strong LD in the data.…”

Prediction of complex human diseases from pathway-focused candidate markers by joint estimation of marker effects: case of chronic fatigue syndrome

“…After that, Malo et al [ 41 ] and Tsai et al [ 42 ] have expressed their own approaches to model covariance between SNP effects in genetic association analyses. Very recently, Fridley and Jenkins [ 43 ] also introduced an approach, which is closely related to our earlier model [ 23 ] for genetic association studies, see also [ 35 , 44 , 45 ] for the other alternatives.…”