Association Analysis of Common Structural Variants in the Alzheimer’s Disease Sequencing Project

Abstract

AbstractBackgroundStructural variants (SV), genomic rearrangements of >50 base pairs, are an important source of genetic variation and have a great impact on gene expression and protein function. However, their contribution to the genetic architecture of Alzheimer’s Disease (AD) has not been comprehensively investigated. Identification of SVs by short‐read alignment can be inaccurate and biased. We used a novel SV calling pipeline that leverages assembly‐based methods and graph‐based representation, to identify SVs with high accuracy in the diverse sample of the Alzheimer’s Disease Sequencing Project (ADSP). We further examined the association of common SVs with AD.MethodsWe applied Biograph for SV calling on the ADSP 17K whole genome sequence data. After filtering out low‐quality SVs and performing sample‐level quality control, we identified 222,956 deletions, 177,139 insertions, and 10,558 inversions in 12,908 individuals (6,304 AD cases; 6,604 controls). We performed association analyses of common SVs (MAF > 1%) with AD using a logistic mixed‐effects model adjusting for sex, technical covariates, principal components, and relatedness. Statistical significance was evaluated using a Bonferroni‐corrected threshold (P< 9.19.x10−7).ResultsThere were 28,634 deletions, 25,522 insertions, and 198 inversions with MAF > 1%. We identified 2 intronic deletions significantly associated with AD. The first deletion is located in CCDC88B, encoding a protein that regulates T‐cell maturation and inflammation. The second deletion is located in CCDC11, predicted to be part of the spliceosomal complex. Among 11 insertions significantly associated with AD, 3 mapped to exonic sequences. A relatively common insertion located in an exon of FOXO6 showed the most significant association with AD (P = 1.41×10−12). FOXO6 is a member of Forkhead transcription factors that have been implicated in aging. Along with its high level of expression in the hippocampus, mice lacking FOXO6 had impaired memory consolidation. We identified one inversion suggestively associated with AD (P = 3.32×10−6) and located in an intron of PDE5A. PDE5A inhibitors have been reported as potential therapeutic targets for AD.ConclusionUsing novel data structures and algorithms to call SV, we discovered 14 SVs significantly associated with AD. Validation and gene prioritization of these findings are warranted.