A New Method For Imaging Genetics Research-Weighted Distance Correlation

Imaging genetics is mainly used to reveal the pathogenesis of neuropsychiatric risk genes and understand the relationship between human brain structure,functional and individual differences.Increasingly,the brain-wide imaging phenotypes in voxels are available to test the association with genetic markers.A challenge with analyzing such data is their high dimensionality and complex relationships.Recently,a class of nonparametric approaches have been carried out to resolve the problem of the correlation analysis between two multivariate variables,with the distance covariance/correlation[1]as the most prominent one.Distance covariance/correlation is introduced to measure both linear and non-linear dependence between two random vectors in arbitrary dimension without relying any model assumption,making it more applicable for processing data in imaging genetics.Yet for the brain-wide imaging phe-notype,the number of voxels could be thousands or even higher.Under such high-dimensional settings,dCor might be invalid.[2]showed the power of the dCor falls sharply as the dimension increases,which implies that dCor is not suitable for brain-wide imaging genetic data.In order to construct a dual multivariate analysis method for phenotype and genetic variations,we introduce a novel framework for association analysis in imaging genetics based on a weighted distance correlation(wdCor)mea-sure.The method are motivated to assign strong weights to the true dependent vox-els and negligible weights to the remaining independent voxels,in this way,the in-dependent coordinates of phenotypes are removed from the test statistic.After that,we propose an adaptive permutation procedure to calculate the p-value of wdCor test.The basic idea is to identify those uncorrelated SNPs at an early stage,so as to reduce the permutation times and calculation burden of GWAS,thus we greatly improve the calculation efficiency.Besides,in extensive simulation studies,wdCor alleviate the power loss of dCor caused by the curse of dimensionality.Finally,We successfully apply wdCor to conduct a large-scale analysis on data from the ADNI.Our method provides new research directions and ideas for double multivariate analysis of high-dimensional data,it can also be used as a tool for scientific analysis of imaging ge-netics research in practical applications.The R package wdcor is available in Github:https://github.com/yangyuhui0129/wdcor...