Genome wide association study gateway—transitioning variants from association to causality in complex diseases

Extract

Complex diseases following common variant-related models are often influenced by multiple genetic variants, each contributing a small effect [1]. Instead of mapping genes, pathways, tissues, or cell types, there are significant challenges when mapping and characterizing a single single nucleotide polymorphism (SNP) as a causal factor in complex diseases, which are influenced by the combined impact of a multitude of different variants rather than just one [2]. The concept of disease-causing variants, when moving from rare Mendelian diseases to common complex diseases, the search for disease-causing variants is shifting its focus from expected inheritance patterns with high penetrance at the phenotype level to consistency of pathogenic mechanisms at the molecular level [3–5].

The required evidence for transitioning variants from association to causality in a complex disease can be summarized into four levels. The first piece of evidence is about the resolution of the association testing at risk loci. Variant typing with high resolution helps in separating potentially causal variants from non-causal proxies in linkage disequilibrium (LD), increasing the likelihood of detecting rare causative variants, large DNA duplications, or deletions with a large impact on gene function [6]. The next piece of evidence comes from functional annotation of variants, such as protein-coding alteration, impact on expression level (eQTL), mRNA splicing (sQTL), regulatory elements, RNA interactions, or chromatin interactions. A credible set of candidate variants can be retrieved using statistical functional fine-mapping methods if multiple annotations and multiple causal variants located in the same loci [7]. The third level of evidence results from assessing the direct biological functional consequences, such as activity and stability of the protein which is affected by coding variant, binding affinity of transcriptional/splicing factors, and expression level of the target gene affected by non-coding variants. These evaluations can be conducted by perturbation experiments under specific cell lines and environmental conditions [8]. The final piece of evidence comes from evaluating the impact of the functional consequences at the phenotypic level, that is, the manifestation of the phenotype and its potential rescue [9].