A scoping review of statistical methods to investigate colocalization between genetic associations and microRNA expression in osteoarthritis

Abstract

Background: Genetic colocalization analysis is a statistical method that evaluates whether two traits (e.g., osteoarthritis [OA] risk and microRNA [miRNA] expression levels) share the same or distinct genetic association signals in a locus typically identified in genome-wide association studies (GWAS). This method is useful for providing insights into the biological relevance of genetic association signals, particularly in intergenic regions, which can help to elucidate disease mechanisms in OA and other complex traits.

Objectives: To review the existing literature on genetic colocalization methods, assess their suitability for studying OA, and investigate their capacity to integrate miRNA data, while bearing in view their statistical assumptions.

Design: We followed scoping review methodology and used Covidence software for data management. Search terms for colocalization, GWAS, and genetic or statistical models were used in the databases MEDLINE and EMBASE, searched till March 4, 2024.

Results: Our search returned 546 peer-reviewed papers, of which 96 were included following title/abstract and full-text screening. Based on both cumulative and annual publication counts, the most cited method for colocalization analysis was coloc. Four papers examined OA-related phenotypes, and none examined miRNA. An approach to colocalization analysis using miRNA was postulated based on further hand-searching.

Conclusions: Colocalization analysis is a largely unexplored method in OA. Many of the approaches to colocalization analysis identified in this review, including the integration of GWAS and miRNA data, may help to elucidate genetic and epigenetic factors implicated in OA and other complex traits.

Keywords: Genetic architecture characterization; Genetic colocalization; Genetic pleiotropy; Genome-wide association study; MicroRNAs; Models; Statistical.

Fig. 1

Schematic illustrating the concept of genetic colocalization analysis between two traits of interest (referred as Trait1 and Trait2). Plot A) illustrates the results (-log₁₀ ​P-value) on the Y axis along genomic positions for Trait1 (top panel) and Trait2 (bottom panel) in a standard GWAS approach for variant discovery. The region highlighted corresponds to a locus on chr16 with variant positions in base pairs (bp) which exhibits association signals for both traits at the genome-wide significance level. This locus can be investigated using colocalization analysis methods to decipher the possible underlying scenarios, as illustrated in plots B and C. Plot B illustrates a possible scenario of genetic colocalization between both traits, where both traits result from the same genetic variant (represented by a single purple star), while Plot C illustrates a possible scenario where the two traits result from distinct variants (as represented by different stars). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 2

Simplified overview of miRNA biogenesis and function in four scenarios showing how SNPs can impact miRNA expression and function. Typically, transcribed miRNA precursors are processed into mature miRNAs, which can lead to mRNA degradation or repression of translation through seed-sequence binding to complementary sequences. Common examples of possible miRNA-SNP interactions are illustrated in scenarios 1–4. SNPs can affect the processing of mature miRNAs or the transcription of miRNA precursors, thereby influencing their overall expression and function (scenarios 1–2). Alternatively, SNPs can alter the binding efficiency between miRNAs and mRNAs, effectively impairing these regulatory functions (scenarios 3–4). Colocalization analysis of OA-associated SNPs corresponding to each scenario, and other miRNA-SNP interactions, can help unravel the underlying miRNA biology that impacts OA pathology.

Fig. 3

PRISMA flow chart detailing the article search and screening process.

Fig. 4

Annual count for each colocalization method by year of publication. Methods that appeared once in the literature search were excluded. Simple Sum (SS) and SS 2 (SS2) were combined. Empirical COnfiguration of Associations with VAriants in R (eCAVIAR); Enrichment Estimation Aided Colocalization Analysis (ENLOC); Pairwise analysis of GWAS (gwas-pw); Hypothesis Prioritisation in multi-trait Colocalization (HyPrColoc); Joint Likelihood Mapping (JLIM); multiple-trait-coloc (moloc); Summary data-based MR/Heterogeneity in Dependent Instruments (SMR/HEIDI).