作者: Derrek P. Hibar , Jason L. Stein , Neda Jahanshad , Omid Kohannim , Arthur W. Toga
DOI: 10.1007/978-3-642-40760-4_75
关键词:
摘要: The SNP-SNP interactome has rarely been explored in the context of neuroimaging genetics mainly due to complexity conducting ~1011 pairwise statistical tests. However, recent advances machine learning, specifically iterative sure independence screening (SIS) method, have enabled analysis datasets where number predictors is much larger than observations. Using an implementation SIS algorithm (called EPISIS), we used exhaustive search genome-wide, identify and prioritize SNPs for interaction analysis. We identified a significant SNP pair, rs1345203 rs1213205, associated with temporal lobe volume. further examined full-brain, voxelwise effects ADNI dataset separately independent healthy twins (QTIM). found that each additional loading epistatic effect was ~5% greater brain regional volume (a protective effect) both QTIM samples.