Twenty years of genome-wide association studies: an essential tool for understanding the human genome

Genome-wide association studies (GWAS) have become an essential tool for identifying genetic variants that influence human traits and complex diseases.

Through comparative analysis of the DNA of large populations, these studies allow scientists to locate small genetic variations associated with diseases such as Alzheimer’s disease, Parkinson’s disease, and diabetes, fostering the development of personalized medicine and new therapeutic strategies.

However, while GWAS have important biomedical applications, they also raise significant bioethical concerns, especially when their findings are used to genetically select embryos or to evaluate complex human traits. This calls for ethical reflection on the limits of these technologies.

GWAS are one of the most influential methodologies in modern genetics. These studies consist of comparative analysis of the entire genomes of a large number of individuals in order to identify statistically significant associations between genetic variants and certain traits or diseases.

The fundamental objective of GWAS is to detect small differences in the DNA sequence that distinguish people who have a certain trait—for example, a disease—from those who do not have it. In other words, the goal is to identify the genetic variants that appear most frequently in individuals who share a specific biological trait, within the vast “sea” of DNA sequences.

This involves analyzing the frequencies of genomic variants in large population samples. Individuals can be classified into two groups: those who have a certain genetic variant and the associated trait (for example, a disease); and those who lack that variant and do not present that trait. Statistical comparison between the two groups allows regions of the genome potentially involved in expression of the trait under study to be identified.

More than 20 years after their introduction into genomic research, GWAS have become a fundamental tool for identifying genetic variants associated with numerous human traits and complex diseases. Their capacity for detection increases notably as the analyzed sample size grows, resulting in the creation of large biobanks and international consortia of genomic data.

Accordingly, they are making a substantial contribution to our understanding of key areas in population genetics, evolutionary genetics, and conservation genetics, and especially in biomedical research for application in personalized medicine, pharmacogenomics, and gene therapy. GWAS also play a role in agriculture and livestock farming for the improvement of cultivated plants and domestic animals. In this article, we will look at the fundamentals of this technology and its applications in functional genomics in man.

Major genes, polygenes, and heritability

To better understand the significance of GWAS results, it may be helpful to review some basic concepts of genetics.

In higher organisms, two types of systems for determining traits can be distinguished: “major” genes and “polygenes”.

Major genes are those that determine well-defined qualitative traits, in which there is a direct relationship between the genotype (the DNA sequence) and the phenotype (the observable manifestation of the trait it encodes). Classic examples are the color of flowers in many plants, the SRY gene involved in the determination of male sex in humans, and the blood groups of the ABO system, etc.

Polygenes, or minor genes, determine traits through the joint action of multiple units distributed throughout the genome, each of which contributes to the additive effect to a varying degree. Polygenic systems are related to quantitative traits such as height, body weight, mental illness, cancer, etc.

In polygenic traits, the phenotype depends not only on genes, but also on environmental factors. Therefore, the total phenotypic variation (V_P) can be expressed as the sum of three components: V_P=V_G+V_E+V_GE, where: V_G represents the variation in the trait due to genes; V_E the variation in the trait due to environmental factors; and V_GE the variation due to the interaction between genes and the environment. From these components, the parameter “heritability” (h²) is calculated; this estimates the proportion of the phenotypic variation of a trait due to genetic factors, where h²=V_G/V_P. It is a central concept in quantitative genetics that allows us to evaluate the relative contribution of genetics and the environment in the manifestation of traits.

The human genome and the molecular basis of genes

The genome is the complete set of genetic information contained in the DNA molecules of an organism. It is made up of one or more DNA molecules, the chromosomes, which in turn consist of tens, hundreds, or thousands of genes and non-coding intergenic regions. The concept of gene has evolved as knowledge about the structure and function of DNA has advanced. Today, we understand that a gene is a DNA sequence that contains the information necessary to produce a functional RNA molecule, which in many cases will be subsequently translated into a protein.

The human genome contains approximately 3.2 billion base pairs, distributed over 23 chromosomes in the haploid set. There are an estimated 21,000 protein-coding genes in this genetic material. Regions of the genome that directly encode proteins are called exons. All of them together constitute the “exome”, which accounts for approximately 1.5% of the human genome. Despite this small proportion, the exome plays a fundamental role in determining the phenotypic traits of each individual. However, the rest of the genome is not a simple set of “junk DNA”, which was the initial assumption after the completion of the Human Genome Project in 2003.

Soon afterwards, thanks to the international ENCODE (ENCcyclopedia Of DNA Elements) project, it would be shown that many of the non-coding regions of the human genome contain regulatory elements that control when, where, and to what extent genes are expressed. ENCODE is a large-scale project involving numerous research groups, which was launched in 2003 by the National Human Genome Research Institute (NHGRI) in the United States with the aim of identifying all the functional elements of the human genome.^[1] The functionality of the entire genome varies from one cell to another, depending on the developmental stage, cell type (specialization), or tissue to which they belong, so that, although they share the same genome, the part that is expressed in each of them varies. The complete set of RNA molecules, especially messenger RNA, present in a cell or tissue at any given time is known as the “transcriptome”.

Fundamentals of GWAS

GWAS are based on the systematic comparison of the DNA of thousands—or even millions—of individuals to detect genetic variations associated with biological traits or diseases.

Most of these analyses focus on single nucleotide polymorphisms (SNPs), which involve the substitution of a single nucleotide base for another (A, T, C, or G) at a specific position in the DNA. An SNP is considered important when it occurs in at least 1% of the population and naturally contributes to phenotypic variation. It is the most frequent type of genetic variation and is mainly found in non-coding DNA. When SNPs appear in a gene or regulatory area of gene expression, they can alter the function or expression of the gene. They are used to locate and map polygenes related to quantitative traits, also called quantitative trait loci (QTLs).

GWAS analysis is based on the development of SNP arrays that enable hundreds of thousands or even millions of these genetic markers to be analyzed simultaneously in each individual DNA sample. These arrays typically contain between 200,000 and over two million SNPs, allowing much of human genetic variability to be explored. In addition to SNPs, there are other forms of genetic variation that can also influence the phenotype. These include “indels”, which are insertions or deletions of nucleotides in the DNA sequence, and “structural variants” (SV), which involve larger changes in the genome, generally greater than 50 base pairs. These variants can play an important role in a number of diseases, including some neurodegenerative disorders.

Once SNPs or other genomic variants have been identified, they are generally used to search for nearby variants that directly contribute to the disease or trait whose genetic basis is under investigation. The analysis consists of determining whether the presence of a certain SNP or any other variation at the same location in the genome is statistically associated with the same phenotype.

Functionality of genetic variants

Analyses derived from the ENCODE project have shown that only about 12% of SNPs are located within coding regions of the genome. In fact, disease-associated SNPs are more likely to be found in non-coding regulatory regions, such as “promoter” or “enhancer” regions, which are usually located outside the gene sequence and are critical for regulating gene expression. This suggests that many diseases are not necessarily due to changes in the protein encoded by a gene, but to alterations in the mechanisms that regulate its expression.

Genomic studies have also revealed that each individual carries more than 10,000 genetic variants—one every 17 nucleotides—of which an estimated 300 may affect protein function. While common variants are often shared among different human populations, many rare variants are characteristic of specific population groups. Therefore, GWAS also have a special interest in population genetics and evolutionary genetics.^[2]

Technological evolution and biomedical applications of GWAS

In the early years of GWAS, so-called “microarrays” were used to detect SNPs and other genetic variations. These systems compared the DNA of individuals to a standard reference sequence in the human genome. However, this approach oversimplified the complexity of genetic variation. In recent years, new mass sequencing technologies have been developed along with the use of multiple reference genomes and more sophisticated bioinformatics tools that allow the direct analysis of individual DNA sequences and more accurate detection of genetic diversity and integration with gene expression data.^[3]

The UK Biobank is a large-scale GWAS database and has contributed significantly to research into the genetic basis of many diseases. Samples in this large biobank come almost entirely from the United Kingdom, which means that GWAS results are based on specific regions and ethnicities. European ancestry is generally overrepresented in GWAS applications, while data based ​​on other regions or ethnicities are more limited.

GWAS have identified thousands of genetic variants associated with complex diseases, including neurodegenerative diseases, schizophrenia, type 2 diabetes (T2D), high blood pressure, and various types of cancer. These studies allow the identification of relevant genes and intergenic interactions, and offer very valuable information about the mechanisms involved in the traits studied and about the underlying mechanisms involved in their expression. These analyses have contributed significantly to the development of personalized medicine, pharmacogenomics, and the design of new therapeutic strategies.

Another fundamental resource for this type of study is the GWAS Catalog. This compendium of GWAS analyses was founded by the NHGRI in 2008, in response to the rapid increase in the number of studies using this technology. The EMBL (European Molecular Biology Laboratory), the world’s leading provider of biological data resources and tools, has also been collaborating since 2010. The Catalog offers an unprecedented opportunity to investigate the impact of common variants on complex diseases. In addition, it greatly facilitates the knowledge and systematic summary of the observed associations. Based on published studies, it is a consistent, searchable, and freely accessible database of SNP-trait associations, which is easily integrated with other resources and accessed by scientists, clinicians, and other users worldwide. It is also one of the largest repositories of full genome-wide summary statistics, which researchers can submit directly.

In recent decades, many research studies have been published on the application of GWAS to determine the genetic basis of several neurodegenerative diseases. These diseases, such as Alzheimer’s disease, Parkinson’s disease, ALS, multiple sclerosis, etc., are characterized by their complex genetic and regulatory patterns. For example, one GWAS study successfully identified approximately 75 genetic risk loci for Alzheimer’s disease dementia,^[4,5] while another identified 90 independent genome-wide significant signals across 78 genomic regions, including 38 novel independent risk signals in 37 loci for Parkinson’s disease.^[6]

Despite these promising results, we are still far from fully understanding the current state of research and future trends in the development of GWAS for these types of diseases. Therefore, in order to better understand their causes, GWAS are being complemented with transcriptome-wide association studies (TWAS), which provide additional insights by allowing the genetic basis of these disorders to be correlated with their functional and physiological aspects.

GWAS face several limitations since, as noted above, they either fail to identify the causal genes, or most of the genetic variants identified are assigned to non-coding regions or have limited power to detect rare variants or small-effect polygenic signals, particularly in small population samples. TWAS studies, on the other hand, integrate tissue-specific QTL data, which makes it easier to locate genetically-regulated gene expression in the tissues in which the disease manifests. Nevertheless, there is still a long way to go since, in the case of Alzheimer’s disease for example, the GWAS carried out so far are either too few or lack samples from non-European populations.

Regarding Parkinson’s disease, in a GWAS published in 2025 that included around 63,000 patients and approximately 1.7 million control samples, a total of 134 risk loci were identified in European, Icelandic, Finnish, and Ashkenazi Jewish populations.^[6] This study also provided a follow-up analysis, aiming to facilitate precision medicine with the aim of advancing research on Parkinson’s disease.

GWAS have been used for many other health-related traits. In the case of T2D, a meta-analysis that included genomic data from more than 2.5 million individuals, of whom 428,452 had T2D, is notable. This study identified 1,289 significant signals in the genome related to an increased risk of this disease, which were mapped to 611 loci.^[7] It also included data on coronary artery disease, peripheral artery disease, and end-stage diabetic nephropathy, and linked obesity and vascular problems to the progression of T2D.

Bioethical assessment

Although GWAS have mostly focused on analyzing disease-related genetic variation, they are equally applicable to any type of quantitative traits governed by polygenic systems. Consequently, they are being applied for traits such as height, IQ, obesity, eye and skin color, etc. Some of these applications are ethically questionable due to their neo-eugenic nature, such as when they are used in preimplantation genetic diagnosis to select certain traits in embryos created through in vitro fertilization. Thus, in a previous article we saw the use of GWAS for analysis of the factors involved in the genetic causes of infertility due to embryonic inviability.^[8,9]  It would also be unethical to apply GWAS for the detection and selection of embryos based on their genetic characteristics related to diseases or other hereditary traits. Regardless of the findings of GWAS or other more or less informative analyses on the genetic traits of the samples, destroying human embryos—whose dignity must be respected at all times—cannot constitute a legitimate means that makes these applications ethically acceptable.

Nicolás Jouve

Professor Emeritus of Genetics at the University of Alcalá

Member of the Bioethics Observatory of the Catholic University of Valencia

 

References

[1] The ENCODE (Encyclopedia of DNA Elements) Project. Science. 306, 5696 (2004): 636-640.

[2] The 1000 Genomes Project Consortium. An integrated map of genetic variation from 1,092 human genomes. Nature 491(2012):56–65.

[3] Genomic Inequality. To successfully use a patient’s genetic makeup in a clinical setting, we must better understand the incredible diversity of human genomes. The Scientist. Dec 1 (2012).

[4] Andrews, S.J., Renton, A.E., Fulton-Howard, B. et al. The complex genetic architecture of Alzheimer’s disease: novel insights and future directions. EBioMedicine 90(2023):104511.

[5] Bellenguez, C., Küçükali, F. Jansen, I.E. et al. New insights into the genetic etiology of Alzheimer’s disease and related dementias. Nat Genet. 54 (2022):412-436.

[6] The Global Parkinson’s Genetic Program, Leonard, H.L.  Novel Parkinson’s Disease Genetic Risk Factors Within and Across European Populations. MedRxiv. DOI:  https://doi.org/10.1101/2025.03.14.24319455

[7] Suzuki K, Hatzikotoulas K, Southam L, Taylor HJ, Yin X, Lorenz KM, et al. Multi-ancestry genome-wide study in >2.5 million individuals reveals heterogeneity in mechanistic pathways of type 2 diabetes and complications. MedRxiv DOI: 10.1101/2023.03.31.23287839

[8] Carioscia, S.A., Biddanda, A., Starostik, M.R. et al. Common variation in meiosis genes shapes human recombination and aneuploidy. Nature (2026). https://doi.org/10.1038/s41586-025-09964-2

[9] JOUVE, N. and TUDELA, J. “A genetic study on embryo inviability reopens the bioethical debate after using 139,000 human embryos” In Bioethics Observatory, Catholic University of Valencia, 2 February 2026.