GWAS Catalog and Genome-Wide Association Resources

  • The NHGRI-EBI Catalog of human genome-wide association studies, commonly known as the GWAS Catalog, is a publicly accessible database that collects, curates, and organizes findings from genome-wide association studies. It is jointly developed by the National Human Genome Research Institute (NHGRI) and the European Bioinformatics Institute (EMBL-EBI). The resource provides information about genetic variants associated with diseases, biological traits, molecular phenotypes, and other measurable characteristics. It is widely used in human genetics, statistical genomics, epidemiology, functional genomics, and precision medicine.
  • A genome-wide association study (GWAS) investigates relationships between genetic variants and observable traits or diseases across a population. Most GWAS examine large numbers of single-nucleotide polymorphisms (SNPs) distributed throughout the genome. Researchers compare allele frequencies or genotype patterns between individuals with different trait values or disease statuses to identify statistical associations. These studies can reveal genomic regions associated with complex conditions such as diabetes, cardiovascular disease, cancer, autoimmune disorders, psychiatric conditions, and many other phenotypes. The GWAS Catalog provides a central resource for discovering and comparing these published findings.
  • The GWAS Catalog was established in 2008 in response to the rapid growth of genome-wide association research. As the number of published studies increased, it became difficult for researchers to locate, compare, and interpret the large volume of reported genetic associations. The Catalog addresses this problem by systematically identifying eligible studies, extracting important information, applying standardized annotations, and making the results available through searchable web pages, downloadable files, visualizations, and programmatic interfaces. Since 2010, its development and delivery have been a collaboration between NHGRI and EMBL-EBI.
  • The main purpose of the GWAS Catalog is to provide a reliable and standardized record of reported variant–trait associations. Its data include information about publications, studies, samples, traits, genetic variants, mapped genes, statistical significance, effect estimates, and population characteristics. The resource also provides access to full genome-wide summary statistics for studies that have made these data available. This distinction is important because the curated association records generally focus on reported significant findings, whereas full summary statistics contain results across the broader set of variants tested in a study.
  • The Catalog identifies eligible GWAS publications through literature searches, including searches of PubMed. Curators examine relevant publications and extract information about study design, sample characteristics, investigated traits, genetic variants, and statistically significant associations. Data are standardized using controlled vocabularies and quality-control procedures. The Catalog also accepts eligible submissions of unpublished GWAS data, allowing researchers to share results before formal publication. New curated data releases are made periodically, so users should record the release or access date when using the resource in research.
  • A GWAS Catalog record can represent several related but distinct entities. A publication describes the scientific article in which results are reported. A study represents a particular analysis or set of analyses within that publication. A sample record describes the individuals included in the analysis, while an association links a genetic variant to a particular trait. A single publication may contain multiple studies, and a study may report many associations. Understanding these distinctions is important when downloading data or interpreting the number of studies and associations returned by a search.
  • Each GWAS Catalog study is assigned a unique accession identifier beginning with the prefix GCST. This identifier makes it possible to refer to a specific study independently of its publication title or PubMed record. Publications may also be identified by PubMed identifiers (PMIDs), while individual variants may be represented using dbSNP RefSNP identifiers, commonly called rsIDs. Genes are identified using standardized gene symbols and database identifiers, and traits are associated with ontology identifiers. These identifiers support cross-referencing between the GWAS Catalog and other biological databases.
  • The Catalog records genetic variants associated with a wide range of phenotypes. These include binary disease outcomes, quantitative measurements, molecular traits, physiological characteristics, behavioural traits, and responses to environmental exposures or medications. Examples include type 2 diabetes, blood pressure, body mass index, cholesterol concentrations, inflammatory markers, gene expression, and drug response. Standardized trait annotation is essential because different publications may use different names for similar phenotypes. A controlled ontology allows related findings to be grouped and compared more consistently.
  • The Experimental Factor Ontology (EFO) is used to standardize many of the traits represented in the GWAS Catalog. EFO provides structured terms for diseases, phenotypes, measurements, and experimental factors. Each standardized trait can be linked to an ontology identifier, helping researchers search for related findings even when the original publications use different terminology. The Catalog also preserves the reported trait, which is the description supplied by the study authors. This distinction is useful because the standardized ontology term supports consistent comparison, while the reported trait preserves details of the original analysis and phenotype definition.
  • A central concept in GWAS analysis is the variant–trait association. An association indicates that a genetic variant is statistically related to a trait in the studied population. The Catalog records information such as the variant identifier, genomic position, mapped gene, risk or effect allele, effect size, p-value, and other relevant study details when available. For binary traits, effect estimates may be expressed as odds ratios, while quantitative traits commonly use beta coefficients. These values help researchers understand the direction and magnitude of an observed association, but they must be interpreted in relation to the study’s phenotype definition, genetic model, allele coding, and statistical analysis.
  • Statistical significance is a major criterion in GWAS. Because hundreds of thousands or millions of variants are tested, conventional significance thresholds such as 0.05 are not sufficiently stringent to control false-positive findings. A widely used genome-wide significance threshold is a p-value of 5 × 10⁻⁸. This threshold accounts approximately for the large number of independent common genetic variants tested in many traditional GWAS settings. However, the appropriate threshold may depend on the study design, number of variants, ancestry composition, variant frequency, and analysis method. Statistical significance also does not establish biological causality.
  • The GWAS Catalog primarily curates eligible reported associations that meet its significance and study-quality criteria. A study may test millions of variants but report only a small subset of statistically significant findings in its publication. The Catalog’s curated association records should therefore not be interpreted as a complete representation of every variant tested in that study. Full genome-wide summary statistics are valuable because they provide results across the tested variants, including associations that do not reach the conventional genome-wide significance threshold. These files allow independent reanalysis, meta-analysis, fine-mapping, and other statistical investigations.
  • The Catalog includes information about study ancestry and population characteristics. Ancestry annotation helps researchers understand which populations contributed to a GWAS and evaluate whether findings are likely to generalize to other groups. Study metadata may include the number of participants, cases and controls, country of recruitment, ancestry labels, and cohort information. These details are important because allele frequencies, linkage disequilibrium patterns, environmental exposures, and disease prevalence can differ across populations. A genetic association discovered in one population may not have the same effect or predictive value in another.
  • Ancestry labels in the GWAS Catalog are standardized using documented population descriptors. These labels help organize studies and facilitate comparisons across datasets. However, genetic ancestry categories are not equivalent to race, ethnicity, nationality, or cultural identity. Many study populations are admixed, and broad ancestry labels may not capture all genetic diversity within a cohort. Researchers should examine the underlying sample descriptions and study design rather than relying exclusively on a simplified population label.
  • One of the most important applications of the GWAS Catalog is the investigation of complex disease genetics. Many common diseases are influenced by numerous genetic variants, each contributing a relatively small proportion of overall risk. GWAS have identified thousands of genomic regions associated with traits such as type 2 diabetes, coronary artery disease, inflammatory bowel disease, asthma, and psychiatric disorders. By collecting these findings, the Catalog helps researchers identify previously reported loci, compare associations across studies, and investigate genetic overlap between different diseases.
  • The Catalog is also useful for studying quantitative traits. These include height, body mass index, blood pressure, blood lipid concentrations, glucose levels, kidney function, and other measurable biological characteristics. Quantitative trait GWAS can identify genetic variants that influence normal biological variation and disease-related processes. Many of these findings help reveal genes, pathways, and molecular mechanisms involved in human physiology. However, an association with a quantitative trait does not necessarily imply that the variant causes a clinical disorder.
  • GWAS Catalog findings are frequently used in functional genomics. A variant associated with a trait may be located within a coding region, intron, promoter, enhancer, or other regulatory region. Many associated variants are non-coding and may influence gene expression rather than directly changing a protein sequence. Researchers therefore integrate GWAS findings with expression quantitative trait loci (eQTLs), chromatin accessibility, transcription-factor binding, epigenomic annotations, and gene expression datasets to investigate possible mechanisms. Such integration helps connect statistical associations with biological processes.
  • The relationship between a GWAS-associated variant and a gene requires careful interpretation. The Catalog may map a variant to a gene based on its genomic location or proximity to a gene. However, the nearest gene is not necessarily the causal gene. Regulatory variants can influence genes located far from the associated position, and a single locus may affect several genes. Functional annotation, molecular QTL evidence, chromatin interaction data, experimental validation, and fine-mapping are often needed to identify the most plausible target gene.
  • Linkage disequilibrium is another important consideration. A GWAS-associated variant may not itself be responsible for the observed trait association. It may instead be correlated with a nearby causal variant because the two alleles are frequently inherited together. This is known as linkage disequilibrium (LD). Researchers use LD reference panels and statistical fine-mapping methods to identify candidate causal variants within associated genomic regions. The GWAS Catalog provides information about reported associations and links to relevant genomic resources, but it does not independently establish that every reported lead variant is causal.
  • Fine-mapping aims to identify the genetic variants most likely to be responsible for an association signal. Methods may use Bayesian statistical models to assign posterior probabilities to candidate variants within a genomic region. A credible set is a group of variants that collectively contains the causal variant with a specified probability under the model. Fine-mapping often requires full summary statistics, linkage disequilibrium information, and appropriate assumptions about the number of causal variants. The Catalog’s curated lead associations are useful starting points, but complete summary statistics are generally more appropriate for detailed fine-mapping.
  • The GWAS Catalog also supports research into pleiotropy, which occurs when a genetic variant or genomic region influences multiple traits. A variant associated with several phenotypes may affect a shared biological pathway or gene, or it may be correlated with different causal variants through linkage disequilibrium. Comparing associations across traits can reveal shared genetic influences and potential biological relationships between diseases. However, apparent pleiotropy must be examined carefully because overlapping samples, correlated phenotypes, and LD can create complex patterns.
  • Another important application is Mendelian randomization (MR). MR uses genetic variants as instrumental variables to investigate potential causal relationships between an exposure and an outcome. For example, genetic variants associated with blood lipid concentrations may be used to investigate whether lipid levels influence cardiovascular disease risk. GWAS Catalog data can help researchers identify candidate exposure-associated variants and locate relevant studies. However, a valid MR analysis requires appropriate instrumental-variable assumptions, including relevance, independence from confounders, and the absence of direct effects on the outcome except through the exposure. The Catalog itself does not establish these assumptions or prove causality.
  • The Catalog provides access to full genome-wide summary statistics when they are available. Summary statistics typically contain variant identifiers, chromosome positions, effect and other alleles, allele frequencies, effect sizes, standard errors, p-values, and sample information. They allow researchers to perform meta-analysis, fine-mapping, genetic correlation analysis, polygenic risk score development, and other downstream investigations. The completeness and format of these files vary by study, and researchers should inspect the documentation, genome assembly, allele coding, and sample characteristics before combining datasets.
  • The GWAS Catalog supports several methods of data access. Users can search through its web interface by publication, study, trait, variant, or gene. Individual pages provide detailed information and allow users to explore associated studies and variants. Data can also be downloaded as spreadsheets or tab-separated files, and full summary statistics are available for eligible studies that provide them. The Catalog’s interactive GWAS Diagram provides a visual overview of significant variant–trait associations across the genome.
  • Programmatic access is particularly useful for large-scale bioinformatics workflows. The GWAS Catalog provides a REST API that allows users to retrieve publications, studies, associations, traits, and variant information in structured JSON format. API queries can be used to search for variants associated with a disease, identify studies of a particular trait, or examine associations involving a gene. Responses are paginated, so scripts must follow the returned pagination links when retrieving large result sets. The currently supported service is REST API version 2; the earlier version 1 API was retired in August 2026. Full genome-wide summary statistics are accessed through the Catalog’s download facilities rather than the REST API.
  • The Catalog’s downloadable files support bulk analysis. Researchers can obtain association data, study metadata, ancestry information, trait mappings, and other supporting files. Bulk downloads are generally more efficient than retrieving thousands of individual records through repeated API calls. For reproducible workflows, the downloaded files should be stored with information about their release, date, source URL, and any filtering or transformation steps applied during analysis.
  • The GWAS Catalog is closely connected to several other genomic and biomedical resources. Ensembl provides genome annotation, gene information, variant consequences, and comparative genomics data. dbSNP supplies identifiers and archived records for genetic variants. gnomAD provides population allele frequencies that are useful for evaluating whether variants are common or rare. The 1000 Genomes Project and other population resources provide genotype and haplotype information for ancestry and linkage disequilibrium analyses. Open Targets Genetics and related tools integrate GWAS and functional genomic evidence to prioritize candidate genes and targets. These resources provide complementary information and should be combined according to the research question.
  • The GWAS Catalog should also be distinguished from clinical variant interpretation databases. ClinVar collects clinical interpretations of genetic variants, while ClinGen evaluates gene–disease relationships and develops expert guidance for variant interpretation. A GWAS association indicates a statistical relationship between a variant and a trait, but it does not automatically mean that the variant is pathogenic or clinically actionable. Most GWAS findings concern complex traits influenced by many genetic and environmental factors. Clinical interpretation therefore requires additional evidence beyond the existence of a Catalog association.
  • Several limitations must be considered when interpreting GWAS Catalog records. Studies differ in sample size, phenotype definition, genotyping platform, imputation panel, statistical method, ancestry composition, and quality-control procedures. Small studies may have limited statistical power, while very large studies may identify associations with extremely small effects. Associations may also be affected by population stratification, sample overlap, winner’s curse, and differences in environmental exposures. The Catalog organizes published evidence, but users must evaluate the quality and limitations of each original study.
  • A further limitation is that the GWAS Catalog focuses on associations reported by eligible studies and does not provide a complete catalogue of every genetic influence on every human trait. Studies that fail to identify significant associations may be less visible in the curated association records. Publication bias and differences in research investment can also influence which traits and populations are represented. The availability of full summary statistics helps address some of these limitations by allowing researchers to investigate results beyond the significant findings highlighted in publications.
  • Reproducibility is essential when using GWAS Catalog data. Researchers should record study accessions, publication identifiers, genome assembly, variant identifiers, effect and other alleles, sample sizes, ancestry information, trait ontology identifiers, and the date of data retrieval. When combining associations from multiple studies, they should examine overlapping cohorts, phenotype definitions, allele harmonization, and statistical methods. Differences in genome assemblies and variant representations can otherwise result in incorrect comparisons or duplicated findings.
  • The NHGRI-EBI GWAS Catalog has become a central resource for organizing and exploring human genetic association findings. It supports research into complex disease genetics, quantitative traits, functional genomics, pleiotropy, fine-mapping, genetic risk prediction, and causal inference. Its value comes from combining systematic literature curation, standardized phenotype annotation, study metadata, variant associations, and access to full summary statistics. When integrated with population variation databases, functional genomics resources, gene–disease validity assessments, and clinical variant databases, it provides an important foundation for modern bioinformatics and precision medicine.
Author:

Leave a Reply

Your email address will not be published. Required fields are marked *