Of the twenty LOAD-associated SNPs discovered by GWAS, each has only a small influence on disease risk (Table 2). This is consistent with other complex diseases. Critics have argued this renders GWAS discoveries largely uninformative for presymptomatic screening, and nonviable for therapeutic intervention. However GWAS is not scattershot; the value of GWAS is in identifying multiple genes that influence a common biological pathways. Genetic risk factors for LOAD cluster into three discrete biological pathways; innate immunity, endocytic vesicle recycling and cholesterol homeostasis \cite{21486313}. The importance of this cannot be overstated; GWAS has enabled researcher to peer beyond the established dogma of the ’Amyloid Cascade Hypothesis’, and nominate novel biological pathways for further exploration. In this sense GWAS is able to instigate a key change in our understanding of disease processes.
Translating GWAS findings ’to the bench’ will be required to understand the mechanistic role of these genes in disease. Pathogenic mutations are typically coding and rare whereas, by design, GWAS identifies common, non-coding changes. For example, of the twenty loci involved in LOAD, all are either intronic or intergenic (Table 2). At each loci, it may be that the tag-SNP is in linkage with one or more functional alleles that are readily ’actionable’ at the bench. For example the intronic CD33 tag-SNPs, rs3865444, is in complete linkage (R^2 = 1) with a splice-site mutation (rs12459419). testitin vitro assays confirm the T allele results in a CD33 isoform lacking exon 2, which encodes the sialic acid binding domain\cite{23946390}. This illustrates that GWAS tag-SNPs can be taken to the bench.
One of the major roadblocks when interpreting GWAS results is the lack of understanding of the non-coding genome. The ENCODE and Epigenomics Roadmap Projects have already begun to demystify some aspects of gene regulation \cite{25693563}\cite{22955616}. Biologist friendly databases such as RegulomeDB and HaploReg allow researchers to identify epigenetic markers that are colocalised with genetic variants, which may be indicative of perturbed regulatory \cite{22955989}\cite{22064851}. Eighty-eight of GWAS SNPs reported to date are either intergenic or intronic, of which 12 percent a located within a transcription factor binding site, and 34 percent are located within a DNase hypersensitivity site \cite{22955616}. Expression quantitative trait loci (eQTL) studies in LOAD do suggest that alleles discovered in GWAS influence gene expression.