Editing GWAS: experimental approaches to dissect and exploit disease-associated genetic variation

Abstract

Genome-wide association studies (GWAS) have uncovered thousands of genetic variants that influence risk for human diseases and traits. Yet understanding the mechanisms by which these genetic variants, mainly noncoding, have an impact on associated diseases and traits remains a significant hurdle. In this review, we discuss emerging experimental approaches that are being applied for functional studies of causal variants and translational advances from GWAS findings to disease prevention and treatment. We highlight the use of genome editing technologies in GWAS functional studies to modify genomic sequences, with proof-of-principle examples. We discuss the challenges in interrogating causal variants, points for consideration in experimental design and interpretation of GWAS locus mechanisms, and the potential for novel therapeutic opportunities. With the accumulation of knowledge of functional genetics, therapeutic genome editing based on GWAS discoveries will become increasingly feasible.

Keywords: CRISPR/Cas; GWAS; Genome editing; High throughput.

Fig. 1

Flow of a typical process from initial GWAS to functional dissection. a A typical GWAS involves selection of the study populations, either case-control cohorts or general populations; genotyping of variants across the genome by single-nucleotide polymorphism (SNP) array or whole genome sequencing; and statistical analysis of variant-trait/disease associations. Regional Manhattan plots (also termed as LocusZoom plots) are generated to show the P values of all variants in a genomic region, to explore the patterns of linkage disequilibrium (LD) between the sentinel variant and each variant, and to annotate the genes within this region. b Statistical fine-mapping and genomic annotations are used to prioritize candidate causal variants. Normally, a credible set of causal variants are prioritized according to posterior inclusion probability (PIP) of each variant and genomic annotations, including chromatin accessibility, histone markers, and transcription factor binding potential, are summarized to guide the following functional studies. c Target genes are predicted according to enhancer-target gene promoter interaction (chromatin confirmation capture) and correlation between causal variant genotypes and target gene expression. ASE, allele-specific expression. d Various experimental approaches are employed to investigate the functions of causal variants and target genes and to link them back to the original phenotype

Fig. 2

The genome editing toolbox. A Repair pathways of DNA double-strand breaks (DSBs) induced by nucleases. B CRISPR-mediated gene interference (CRISPRi) or gene activation (CRISPRa) regulates the activity of CREs by sterically blocking or changing epigenetic modifications. C Base editors used a fused deaminase to catalyze the conversion of C to U (C base editor) or A to I (A base editor) on one strand, followed by DNA repair on the non-edited strand. Eventually, the original C:G (or A:T) is converted to T:A (G:C) during DNA replication. D Primer editing leverages nCas9 to nick the genomic target, after which the reverse transcriptase copies the sequence information from the prime editing guide RNA (pegRNA) and replaces the original sequence at the target site

Fig. 3

Genome editing strategies for interrogation of GWAS loci. Approaches used to investigate the functions of causal variants included introduction of indels nearby causal variants (to disrupt the core sequence of a putative transcription factor binding motif), deletion of the genomic region surrounding causal variants, CRISPRi/a (to repress or enhance the activity of local CREs), and allele substitution to change causal variants from one allele to the other (to precisely mimic the genotype). Multiplex CRISPR is used to target multiple causal variants where they can function jointly or synergistically. CRISPR/Cas screening can enable high-throughput interrogation of causal variants or target genes. Target gene knockout is typically achieved by introducing frameshifts into the coding sequence