Humanomni1 quad beadchip

All subjects were genotyped using the Illumina HumanOmni1-Quad BeadChip. Table 1 summarizes the number of individuals in different categories. After quality control, 2024 samples were retained for the final analysis. Details on the study population characteristics and the quality control parameters applied on the samples and markers can be found elsewhere [17] (link), [18] . Briefly, samples were filtered if they had a missing genotype rate ≥0.02, or with mean X-chromosome heterozygosity ≥0.02 for males or outside the range of 0.25∼0.4 for females. Genetic outliers and those with evidence of relatedness (IBD estimate ≥0.4) or non-European ancestry based on genotype data were also excluded. After filtering, 496 AD+P cases, 639 AD intermediate P, 156 AD−P cases and 958 unaffected controls were retained. Markers were excluded if they had a genotype missing rate of >0.02. Markers were examined to determine if missing depended on case/control status or the genotyping batch.

Of the 607 samples in our DNAm dataset, 590 had whole-genome genotyping data available, which was measured using the Illumina HumanOmni1-Quad BeadChip. Standard QC measures were applied: removing samples with less than 95% SNP call rate, sex discrepancies, relatedness (Pi-hat > 0.2), or excess hetero- or homozygosity. Markers with less than a 98.5% call rate, or are monomorphic were removed. Phasing was performed using SHAPEIT [25 (link)] followed by SNP imputation via the IMPUTE2 software [26 (link)], and all individuals in the 1000 Genomes Project as a reference sample. Genetic ancestry was determined using EigenStrat program [21 (link)] and eigenvectors were utilized in statistical analyses, as described in detail below. Given our interest in the role of SNPs in producing gap signals, we limited all of our analyses to the 590 samples that had both genotype and 450k data. We also limited our analysis to those SNPs with a minor allele frequency ≥0.5%, as this value corresponded to the same number of individuals that would be allowed in the smallest gap signal group according to the default input arguments (see “gap hunting” algorithm section).

To represent the characteristics of a typical GWAS panel, markers from the Affymetrix Genome-Wide Human SNP Array 6.0, the Illumina Human1M–Duo Genotyping BeadChip, and the Illumina HumanOmni1-Quad BeadChip were downloaded from the UCSC genome browser, using the table browser tool [24] . The union of these three arrays consisted of 1,936,864 unique SNPs from the 22 autosomes. Because of its unique LD and genic properties, the MHC region (chr6: 29624809–33160245 on build 37) was excluded from downstream analyses.
LD proxies or “tagging” SNPs (r²> = 0.8) for the GWAS panel SNPs were identified using VCFtools [25] (link) based on data from the (N = 379) Europeans (Phase I, version 3, March 14, 2012) in the 1000 Genomes Project [26] (link).
GWAS “non-hits” were defined as all those SNPs in our union GWAS set which were neither a GWAS “hit” (see below), nor in high LD (r²> = 0.8) with a GWAS hit.

Germline DNA was extracted from EDTA-venous blood samples using FlexiGene DNA Kit (Qiagen). Genotyping was performed by the SNP&SEQ Technology Platform, Uppsala, Sweden, using the HumanOmni1-Quad beadchip (Illumina). Five SNPs of interest (rs11979158, rs2252586, rs4295627, rs55705857, and rs78378222) were not represented on the chip and were therefore imputed using the software IMPUTE2 with data from the 1000 Genomes Project as the reference population [17 (link)]. The imputation info scores for the imputed SNPs were 0.997, 0.939, 0.999, 0.535, and 0.795, respectively, where values near 1 indicate that a SNP has been imputed with high certainty. Due to the low imputation info score (<0.85), rs55705857 (8q24.21) and rs78378222 (in TP53, 17p13.1) were excluded from further analysis.