Premium
Analysis of individual families implicates noncoding DNA variation and multiple biological pathways in Alzheimer’s disease risk
Author(s) -
Wijsman Ellen M.,
Day Tyler R.,
Thornton Timothy A.,
Horimoto Andrea R.,
Blue Elizabeth E.,
Bis Josh C.,
Sohi Harkirat K.,
Nato Alejandro Q.,
Nafikov Rafael A.,
Navas Patrick,
Saad Mohamad,
Tsuang Debby W.,
Barral Sandra,
Vardarajan Badri N.,
Beecham Gary W.,
Martin Eden R.,
van Duijn Cornelia M.,
PericakVance Margaret A.,
Mayeux Richard
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.046456
Subject(s) - genetics , pedigree chart , biology , gene , genetic association , genetic linkage , population , computational biology , allele , single nucleotide polymorphism , genotype , medicine , environmental health
Background Late‐onset Alzheimer’s disease (AD) is a complex disorder with multiple genetic risk factors. Linkage and association analysis have mapped dozens of loci in pooled analysis of many pedigrees or large numbers of unrelated cases and controls. Identification of the underlying DNA risk variants in the regions of interest (ROIs) has been complicated by both the genetic heterogeneity and the cost, until recently, of comprehensive DNA sequencing in ROIs. The known loci also leave much heritability unexplained. Method We used the families in the AD Sequencing Project (ADSP) discovery family sample to identify variants of interest from whole genome sequences (WGS), and through the variants, genes implicated in risk. We used SNP‐based multipoint linkage analysis to identify ROIs with rare VOIs, carrying out analysis without trimming pedigrees. We pursued all ROIs with family‐specific lod max scores >1.9, reducing the variants of interest by several filters. We carried out pedigree‐based genotype imputation from the available WGS data, followed by family‐based association analysis, filtered for low population minor allele frequency. We prioritized genes with a low false‐discovery rate for association of single‐cell transcription in brain with AD disease state (PMID:31209304), and genes with high expression in bulk brain (PMID: 24309898). Result We obtained 46 distinct ROIs representing lod max 1.9‐3.5 per ROI in each of 26 of the 110 ADSP discovery families analyzed. 29 ROIs further investigated in 16 of the families yielded 59 prioritized genes, with 1‐11 genes/ROI. Only 4 out of 321 variants that passed all filters in these genes were in exons, with minimal overlap with genes identified in AD GWASs. Only one ROI occurred in two families, with evidence for a shared‐haplotype between these families, implicating FBXO2 and FBXO44 . Both genes are implicated in ubiquitination, while FBXO2 interacts with BACE1 . Multiple pathways, both known and new, are implicated, including the ubiquitin‐proteasome system, neural development and maintenance, and mitochondrial functions. Conclusion This analysis underscores the evidence for extensive genetic heterogeneity and rare variants underlying AD risk, along with multiple potential mechanisms. The preponderance of prioritize non‐coding variants suggests alterations in gene regulation and/or expression as an aspect of AD genetic risk.