Search

Genetic studies have shifted to sequencing-based rare variants discovery after decades of success in identifying common disease variants by Genome-Wide Association Studies using Single Nucleotide Polymorphism chips. Sequencing-based studies require large sample sizes for statistical power and therefore often inadvertently introduce batch effects because samples are typically collected, processed, and sequenced at multiple centers. Conventionally, batch effects are first detected and visualized using Principal Components Analysis and then controlled by including batch covariates in the disease association models. For sequencing-based genetic studies, because all variants included in the association analyses have passed sequencing-related quality control measures, this conventional approach treats every variant as equal and ignores the substantial differences still remaining in variant qualities and characteristics such as genotype quality scores, alternative allele fractions (fraction of reads supporting alternative allele at a variant position) and sequencing depths. In the Alzheimer's Disease Sequencing Project (ADSP) exome dataset of 9,904 cases and controls, we discovered hidden variant-level differences between sample batches of three sequencing centers and two exome capture kits. Although sequencing centers were included as a covariate in our association models, we observed differences at the variant level in genotype quality and alternative allele fraction between samples processed by different exome capture kits that significantly impacted both the confidence of variant detection and the identification of disease-associated variants. Furthermore, we found that a subset of top disease-risk variants came exclusively from samples processed by one exome capture kit that was more effective at capturing the alternative alleles compared to the other kit. Our findings highlight the importance of additional variant-level quality control for large sequencing-based genetic studies. More importantly, we demonstrate that automatically filtering out variants with batch differences may lead to false negatives if the batch discordances come largely from quality differences and if the batch-specific variants have better quality.

Leucine rich repeat transmembrane protein 3 (LRRTM3) is member of a synaptic protein family. LRRTM3 is a nested gene within α-T catenin (CTNNA3) and resides at the linkage peak for late-onset Alzheimer's disease (LOAD) risk and plasma amyloid β (Aβ) levels. In-vitro knock-down of LRRTM3 was previously shown to decrease secreted Aβ, although the mechanism of this is unclear. In SH-SY5Y cells overexpressing APP and transiently transfected with LRRTM3 alone or with BACE1, we showed that LRRTM3 co-localizes with both APP and BACE1 in early endosomes, where BACE1 processing of APP occurs. Additionally, LRRTM3 co-localizes with APP in primary neuronal cultures from Tg2576 mice transduced with LRRTM3-expressing adeno-associated virus. Moreover, LRRTM3 co-immunoprecipitates with both endogenous APP and overexpressed BACE1, in HEK293T cells transfected with LRRTM3. SH-SY5Y cells with knock-down of LRRTM3 had lower BACE1 and higher CTNNA3 mRNA levels, but no change in APP. Brain mRNA levels of LRRTM3 showed significant correlations with BACE1, CTNNA3 and APP in ∼400 humans, but not in LRRTM3 knock-out mice. Finally, we assessed 69 single nucleotide polymorphisms (SNPs) within and flanking LRRTM3 in 1,567 LOADs and 2,082 controls and identified 8 SNPs within a linkage disequilibrium block encompassing 5'UTR-Intron 1 of LRRTM3 that formed multilocus genotypes (MLG) with suggestive global association with LOAD risk (p = 0.06), and significant individual MLGs. These 8 SNPs were genotyped in an independent series (1,258 LOADs and 718 controls) and had significant global and individual MLG associations in the combined dataset (p = 0.02-0.05). Collectively, these results suggest that protein interactions between LRRTM3, APP and BACE1, as well as complex associations between mRNA levels of LRRTM3, CTNNA3, APP and BACE1 in humans might influence APP metabolism and ultimately risk of AD.

BACKGROUND:The insulin-degrading enzyme gene (IDE) is a strong functional and positional candidate for late onset Alzheimer's disease (LOAD). METHODOLOGY/PRINCIPAL FINDINGS:We examined conserved regions of IDE and its 10 kb flanks in 269 AD cases and 252 controls thereby identifying 17 putative functional polymorphisms. These variants formed eleven haplotypes that were tagged with ten variants. Four of these showed significant association with IDE transcript levels in samples from 194 LOAD cerebella. The strongest, rs6583817, which has not previously been reported, showed unequivocal association (p = 1.5x10(-8), fold-increase = 2.12,); the eleven haplotypes were also significantly associated with transcript levels (global p = 0.003). Using an in vitro dual luciferase reporter assay, we found that rs6583817 increases reporter gene expression in Be(2)-C (p = 0.006) and HepG2 (p = 0.02) cell lines. Furthermore, using data from a recent genome-wide association study of two Croatian isolated populations (n = 1,879), we identified a proxy for rs6583817 that associated significantly with decreased plasma Abeta40 levels (ss = -0.124, p = 0.011) and total measured plasma Abeta levels (b = -0.130, p = 0.009). Finally, rs6583817 was associated with decreased risk of LOAD in 3,891 AD cases and 3,605 controls. (OR = 0.87, p = 0.03), and the eleven IDE haplotypes (global p = 0.02) also showed significant association. CONCLUSIONS:Thus, a previously unreported variant unequivocally associated with increased IDE expression was also associated with reduced plasma Abeta40 and decreased LOAD susceptibility. Genetic association between LOAD and IDE has been difficult to replicate. Our findings suggest that targeted testing of expression SNPs (eSNPs) strongly associated with altered transcript levels in autopsy brain samples may be a powerful way to identify genetic associations with LOAD that would otherwise be difficult to detect.