Development of a massively parallel reporter assay to identify functional regulatory variants in the PICALM Alzheimer disease associated locus

Abstract

AbstractBackgroundMost significant genome‐wide association markers for Alzheimer disease (AD) are in noncoding portions of the genome making assigning a corresponding gene involved in risk challenging. Coding variants identified in some these GWAS loci thus far do not explain the proportion of risk attributable to the associated markers, indicating a contribution of noncoding variants to AD. We set out to develop a high throughput method investigating the functional role of noncoding variants associated with AD using the PICALM locus on chromosome 14:85,600,000‐85,860,000, which has been associated with AD across multiple ancestries, as a proof of concept region.MethodWe used the gnomAD database of over 15,000 whole genomes to identify all variants located in the PICALM locus across a wide variety of ancestral populations (non‐Hispanic whites, Hispanics, African Americans, and Asians). We developed a pipeline for epigenomic annotation from available databases and the subsequent design of oligos carrying wild‐type and variant alleles for each variant. For unbiased analyses of functional potential to regulate enhancer activity of each of the alleles we established a large vector library for massively parallel reporter assays (MPRA) in AD relevant cell lines.ResultWe identified >23,000 variants for analysis in the 260 kb PICALM locus. More than 4,000 of those had prior evidence for location in transcription factor binding sites, putative promotors/enhancers and eQTLs. 180 bp oligos centered around the variant alleles were synthesized by TWIST Biosciences. 20 bp random barcodes were added to each of the oligos through PCR amplification with a barcoded primer library. Sequencing of the resulting vector library indicated highly efficient incorporation with 95% unique barcodes present in the library. Reporter gene GFP and a minimal promoter were subsequently cloned in between the oligo and the barcode to establish the final MPRA library for transformation in cell lines.ConclusionHere we established a computational framework for in‐silico epigenomic annotation and oligo design for subsequent high‐throughput analyses and developed an efficient protocol to create a final MPRA library with highly unique barcodes at fragment incorporation. Cell line transformations and transcriptome analyses of the PICALM library to identify regulatory variants are ongoing.