A structured brain‐wide and genome‐wide association study using ADNI PET images

Background Detecting association signals between the brain and genome using ultrahigh-dimensional PET neuroimaging scans GWAS SNP data Method A multi-stage variable selection method is introduced for detecting in structured brain-wide genome-wide studies (brain-GWAS). Validity of proposed approach demonstrated by both theoretical investigation numerical simulations. We apply our to a brain-GWAS ADNI imaging genomic data. Result confirm previously reported (e.g. IL8RB COLEC12 genes) also uncover several novel SNPs genes GALNT4 RIN2 that either are associated with glucose metabolism or have their significantly modified Alzheimer's disease status. Conclusion Compared conventional single-voxel-to-single-SNP methods, more efficient powerful selecting important integrating anatomic gene grouping structures genome, respectively. It avoids resorting large number multiple comparisons while effectively controlling false discoveries.