Genetics and statistical study design · BEGINNER LESSON
Genome-wide association study (GWAS)
A genome-wide association study, or GWAS, is a study that scans many genetic variants across the genome and tests whether any are statistically associated with a trait. It measures differences in variant frequencies between people or groups with different phenotypes, where a phenotype is an observed trait or disease feature.
What it is and what is measured
A GWAS compares genetic variation across the genome to find variants associated with a trait, such as a disease or measurable characteristic. In this setting, a variant is a change at a DNA position, and a single-nucleotide polymorphism (SNP) is a common one-letter DNA difference. The main measurement is statistical association: whether the frequency of a variant differs in a way linked to the phenotype being studied. GWAS often use covariates, which are extra variables such as age, sex, and principal components, to reduce misleading associations.
[1] Monitoring diversity in genome-wide association studies ... [3] PP-GWAS: Privacy Preserving Multi-Site Genome-wide Association Studies
How it works
GWAS studies test each variant against the phenotype, often using models that control for confounding, meaning outside factors that can distort results. Principal components, which summarize broad genetic ancestry, can help avoid false positives caused by population stratification, where ancestry differences make variants look associated even when they are not. Some GWAS are done in a centralized way, while others use distributed or privacy-preserving designs across multiple sites. Statistical thresholds are used to decide which associations are strong enough to report.
[1] Monitoring diversity in genome-wide association studies ... [3] PP-GWAS: Privacy Preserving Multi-Site Genome-wide Association Studies
What affects results
GWAS results can be influenced by phenotype definitions, environmental exposures, technical platforms, and the way samples are grouped by country or ancestry. One source notes that broad ancestry categories and country of recruitment alone do not fully capture heterogeneity, especially immigration-related factors. Another source notes that GWAS can identify loci, but not always the exact relevant gene, because many associated variants are non-coding and may not map cleanly to a gene with a clear effect.
[1] Monitoring diversity in genome-wide association studies ... [2] Specificity, length and luck drive gene rankings in association studies [3] PP-GWAS: Privacy Preserving Multi-Site Genome-wide Association Studies
Limitations
GWAS findings can be hard to compare across studies if data collection methods differ. When studies are combined, assumptions about equal sample sizes across countries can distort diversity estimates, and the same sample may be counted more than once in overlapping meta-analyses. GWAS also do not directly rank genes by trait importance, and rare coding-variant methods such as loss-of-function burden tests can prioritize different genes than GWAS. In short, GWAS is powerful, but its conclusions depend on study design, sample structure, and the kinds of variants being tested.
[1] Monitoring diversity in genome-wide association studies ... [2] Specificity, length and luck drive gene rankings in association studies
SEE THE IDEA
Basic GWAS workflow
Genome-wide variants such as SNPs are collected from many people.
A measurable trait, such as disease status or a quantitative value, is recorded.
Each variant is tested for statistical association with the phenotype, often with covariates.
Significant variants are reported, while ancestry, batch, and other confounders are considered.
Data from people and their traits are compared step by step to find variants linked to the trait.
GWAS tests many variants across the genome, then filters results with statistical and design checks.
TRY IT YOURSELF
Test your understanding
Choose an answer, then check the explanation.
PUT IT TOGETHER
Match each concept to its meaning
Drag a label onto its description, or select a label and then select a description. This also works with a keyboard.
4 pairs to match
Sources and further reading
This general lesson was generated from public source material. It explains a concept; it does not interpret a private paper or provide medical advice.