Editorial: Advancements and Prospects of Genome-wide Association Studies

Ayo Priscille Doumatey ^1* Yafang Li ² Juan Carlos Fernandez-Lopez ³

• ¹ National Human Genome Research Institute, National Institutes of Health (NIH), Bethesda, Maryland, United States
• ² Baylor College of Medicine, Houston, Texas, United States
• ³ National Institute of Genomic Medicine (INMEGEN), Mexico City, México, Mexico

Since the very first Genome-Wide association study was published in 2005 followed by the pivotal manuscript by the Wellcome Trust Case Control Consortium (WTCCC) in 2007 (Klein et al., 2005;Wellcome Trust Case Control, 2007), several thousands of Genome-Wide association studies (GWAS) have been conducted (>45 000) and yielded a myriad of loci associated with diverse human phenotypic traits or diseases (>5000) (Sollis et al., 2023). These studies have not only shed light on the involvement of common variants but have also discovered important insights into biology (Uffelmann, 2021) The field of GWAS has seen tremendous advances both technologically and methodologically. For example, GWAS have been facilitated by single nucleotide polymorphisms (SNPs)-arrays containing a couple of thousand hundred to 2 million SNPs (Visscher et al., 2017). The ability to impute untyped SNPs has increased marker density, improved the statistical power of GWAS, and enabled large-scale metaanalyses across populations, but also aided in fine-mapping regions of interest in genetic loci associated with diseases or phenotypic traits (Cahoon et al., 2024;Zhihui Zhang, 2022). Additionally, with the advances in sequencing technologies and the drop in sequencing cost, the latest GWAS and follow-up studies have been sequence-based and allowed for the assessment of low-frequency and rare variants in association studies (Acar et al., 2023;McMahon et al., 2021;Pan et al., 2023). Despite all the advances in Genomics, not all global populations are evenly represented in GWAS. GWAS have been mainly limited to populations of European descent and could exacerbate existing health disparities and the implementation of precision medicine (Doumatey et al., 2023). The push for genomic equity and representation of all global populations in biomedical research has led to initiatives such as the1000 Genomes, the Human Heredity& Health (H3Africa), and Trans-omics for Precision Medicine (TOPMed), and All of Us in recent years (Auton et al., 2015;Investigators, 2024;Peprah et al., 2017;Taliun et al., 2021). Some of these studies have improved imputation references, mainly 1000 genomes and TOPMed (Kowalski et al., 2019). Others, such as H3Africa, have helped develop population-specific genotype arrays, thus improving discovery in these populations. In this research topic, Brandenburg, JT et al. used such custom array enriched for African common variants, the H3A custom genotyping array, in the largest GWAS for urinary albumin-to-creatinine ratio (UACR), a biomarker of kidney disease, in Sub Saharan Africans living in different regions of the continent and non-resident individuals of African ancestry (UK BioBank) and African American cohorts (CKD-Gen African Ancestry). The authors identified two novel variants associated with UACR in populations of African ancestry (BMP6, HBB), and they replicated only three out of 60 known UACR-associated loci previously identified in European and multiancestry studies. Additionally, the authors performed polygenic score (PGS) and comparison analyses to evaluate the transferability of PGS derived from non-African and multi-ancestry populations to African populations. https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2024.1372042/full).A comprehensive review by Bruner, S.W. et al, in this research topic underscored the need to account for population-biased findings by improving representation across various ancestral populations, which should, in turn, address replication and transferability challenges across populations. The authors especially highlighted advancements in methods and techniques and available resources to conduct functional annotations of GWAS variants and mechanistic "variant-to-function follow-up studies. They described functional annotations (e.g. epigenetic annotations, pathway and network analyses, quantitative trait loci) and functional validations (reporter assays, genome editing, and animal models) as approaches to characterize identified GWAS variants and associated effector genes. https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2024.1375481/full. Some of these methods were used in the studies included in this research topic. In studies conducted by Abbas, M et al., and Brandenburg, JT et al. in populations of African ancestries, the authors leveraged expression quantitative trait locus (eQTL) analyses to understand the mechanisms connecting identified GWAS variants to phenotypes (ref. Abbas et al., Brandenburg et al.) using two different strategies. Brandenburg, JT et al first conducted a GWAS and subsequently carried out a functional annotation of the lead single nucleotide polymorphisms (SNP) associated with UACR using relevant tissue expression profile (glomerular and tubulointerstitial tissues) for a cis-eQTL analysis. In contrast, Abbas, M et al first conducted a differential expression analysis to identify mRNAs that were differentially expressed between Low-Density Lipoproteins, LDL groups (low tertile vs upper tertile), then extracted all SNP in the cis region of each mRNA differentially expressed between LDL groups to perform an eQTL analysis. eQTLs significantly associated with differentially expressed mRNAs were overlapped with genetic variants previously associated with LDL and LDL traits in the GWAS catalogue database. The authors also used pathway and network analyses to further decipher functional relationships between transcriptomics, genomics, and phenotype. Noteworthy, an African-specific eQTL was associated with TTC38 expression in African Americans. TTC38 was significantly upregulated in individuals with elevated LDL. https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2024.1345541/full.Animal models are proposed as a variant-to-function method to understand GWAS discovery. These methods can take many years but effectively decipher the underpinnings of biology in humans. In one of the studies, Fowler S et al. conducted a GWAS in heterogeneous stock (HS) to identify genomic regions associated with intraocular pressure (IOP). The authors indicated a few advantages for conducting GWAS and follow-up studies in rats for phenotypes like IOP: 1) the ability to control the environmental variables known to affect IOP in humans, 2) rats and humans share similar anatomical and developmental characteristics of the eyes (e.g. the aqueous outflow pathway), thus an ideal model to study pathophysiology of ophthalmologic disorders, 3) tissues for phenotyping and gene expression can be easily collected. The combination of GWAS in ~1800 HS rats identified 5 candidate genes, of which two were novel (Ctsc2 and Plekhf2-never reported in human studies); the authors concluded the use of HS rats resulted in new findings that could provide new insights into the molecular basis of IOP. https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2022.1029058/full GWAS can identify genetic variants linked to a trait/disease, but they cannot always establish causality. Thus, epidemiology-based methods such as Mendelian randomization (MR) have been used to establish causality between genetic variants, used as instrumental variables, risk factors and lifestyle factors of several diseases (Benn & Nordestgaard, 2018). Lin, X et al. used this strategy to establish a causal relationship between several reproductive factors (e.g. age at menopause, age at menarche, age at first live birth) and bone density. Their findings suggest that early menopause and late childbirth may be