Bioinformatics strategy to advance the interpretation of Alzheimer's disease GWAS discoveries: The roads from association to causation

Abstract Introduction Genome-wide association studies (GWAS) discovered multiple late-onset Alzheimer's disease (LOAD)-associated SNPs and inferred the genes based on proximity; however, the actual causal genes are yet to be identified. Methods We defined LOAD-GWAS regions by the most significantly associated SNP ±0.5 Mb and developed a bioinformatics pipeline that uses and integrates chromatin state segmentation track to map active enhancers and virtual 4C software to visualize interactions between active enhancers and gene promoters. We augmented our pipeline with biomedical and functional information. Results We applied the bioinformatics pipeline using three ∼1 Mb LOAD-GWAS loci: BIN1 , PICALM, CELF1 . These loci contain 10–24 genes, an average of 106 active enhancers and 80 CTCF sites. Our strategy identified all genes corresponding to the promoters that interact with the active enhancer that is closest to the LOAD-GWAS-SNP and generated a shorter list of prioritized candidate LOAD genes (5-14/loci), feasible for post-GWAS investigations of causality. Discussion Interpretation of LOAD-GWAS discoveries requires the integration of brain-specific functional genomic data sets and information related to regulatory activity.