Biology Atlas

Map / Assays

assay · intermediate · partly written

Genome-wide association study (GWAS)

A GWAS compares the frequency of many common variants between large groups of people (cases and controls, or high and low values of a trait) to flag regions of the genome associated with the trait.

Also called: GWAS

What a GWAS is and is not

A GWAS compares many people, variant by variant. Analysing one patient's variants against the reference and population databases is a different activity; his sentence runs the two together. 1

Reported, not yet replicated

Design and threshold more

Risch and Merikangas argued in 1996 that for common disease with small effects, association studies in many unrelated people are far more powerful than family linkage, which would need over about 2,500 families. The conventional threshold of P below 5 x 10^-8 follows the logic of about 10^6 independent tests (0.05 divided by 10^6); how it became the standard was not checked. 1

Reported, not yet replicated

A locus, not a gene more

The first clear success, in age-related macular degeneration (96 cases, 50 controls, 116,204 SNPs), found an intronic CFH association; the coding change came from following the locus. Across the field, over 80% of associations lie outside coding regions, and the GWAS population is 82 to 86% European in ancestry, which limits where results transfer. 234

Reported, not yet replicated

Who gets measured deep

Backman and colleagues' UK Biobank exome analysis is about 95% European ancestry, the same skew. 5

Reported, not yet replicated

Question answered

Which common variants are more frequent in people with a trait than in people without it. It compares many people at once, one variant at a time. It is not the analysis of one patient's variants: that is a different activity with different output (a short list and a class, not an association). 1

Reported, not yet replicated

Readout physical

Genotypes at hundreds of thousands to millions of positions per person, read from a genotyping array (tag SNPs) or from sequencing, then a test of allele frequency difference per variant. 2

Reported, not yet replicated

Not measured

Which variant is causal, which gene it acts on, and in which direction. Most associated variants lie outside protein-coding regions, and the associated variant is a stand-in for correlated neighbours (linkage disequilibrium). 32

Reported, not yet replicated

Setting

Where in cascade

ML role

Does

Machine-learning models sit after the GWAS, ranking genes at a locus (the locus-to-gene model combines distance, expression colocalisation, variant effect and chromatin features) and predicting variant effect. At a score of 0.5 that model reached precision 0.73 and recall 0.53 against 0.56 and 0.60 for the closest gene, on a gold standard the authors say favours near genes. 6

Reported, not yet replicated

Cannot

It cannot turn an association into evidence of causation or of the direction of a drug effect; 43.2% of loci received no gene at the 0.5 score. 6

Reported, not yet replicated

Still open here

Not yet written: Sample prep, Conditions, Throughput, Time per run, Cost, Engineerable variants, Use cases, Before after, Capital, Data availability.

Linked from

Sources

  1. Risch, N., Merikangas, K.. The Future of Genetic Studies of Complex Human Diseases. Science (1996). paper; opened 2026-11-17 all records citing this
  2. Klein, R., Zeiss, C., Chew, E., et al.. Complement Factor H Polymorphism in Age-Related Macular Degeneration. Science (2005). paper; opened 2026-11-17 all records citing this
  3. Manolio TA, Collins FS, Cox NJ, et al.. Finding the missing heritability of complex diseases. Nature (2009). paper; opened 2026-11-17 all records citing this
  4. Abdellaoui, A., et al.. 15 years of GWAS discovery: realizing the promise. American Journal of Human Genetics (2023). paper; opened 2026-11-17 all records citing this
  5. Backman JD, Li AH, Marcketta A, et al.. Exome sequencing and analysis of 454,787 UK Biobank participants. Nature (2021). paper; opened 2026-11-17 all records citing this
  6. Mountjoy E, Schmidt EM, Carmona M, et al.. An open approach to systematically prioritize causal variants and genes at all published human GWAS trait-associated loci. Nat Genet (2021). paper; opened 2026-11-17 all records citing this
Where it sits on the map
History of this record (2)

This record has been changed 2 times, on the dates below. The note kept against each change is written for the people who maintain the atlas, in their own shorthand, so it is not shown here; it is in the record itself.

Throughput class
high
Emerging
no
Status
partly written
Last updated
2026-11-17
Record
content/assays/genome-wide-association-study.json