Linkage and genetic association studies (GWAS)
- Study design
They look for the region or variant of the genome that travels with a disease: in families (linkage) or comparing cases and controls (association).
What it is
Linkage studies follow, in families with several affected members, which genetic markers are inherited together with the disease; they work best for genes of large effect. Association studies compare the frequency of variants between unrelated cases and controls: first with candidate genes chosen by hypothesis, and today with genome-wide association studies (GWAS), which test hundreds of thousands to millions of single nucleotide polymorphisms (SNPs). Because so many tests are run, they need a very strict significance threshold and replication.
How it is built
- Cases and controls are genotyped on a SNP chip; genotyping quality is checked.
- Stratification (ancestry) is corrected with principal components or controls from the same population.
- Each SNP is tested separately; results are shown in a Manhattan plot.
- Findings are confirmed in an independent replication sample.
Worked example
- GWAS with 2,000 cases and 2,000 controls (4,000 alleles per group). 1,000,000 SNPs are tested.
- SNP X: risk allele in 1,200 case alleles (30%) and 1,000 control alleles (25%).
- Allelic OR = (1,200 × 3,000) / (2,800 × 1,000) = 1.29.
- Chi-square ≈ 25.1, with p ≈ 5.5 × 10⁻⁷: significant for a single test.
- Genome-wide threshold: 0.05 / 1,000,000 = 5 × 10⁻⁸. The p of X is about 11 times larger.
- Reading: X is promising but does not reach genome-wide significance; it needs replication in another sample.
Linkage, candidate gene and GWAS
| Approach | Who | What it detects best | Main limitation |
|---|---|---|---|
| Linkage | Families with several affected members | Rare genes of large effect | Broad regions; little power in complex diseases |
| Candidate gene | Cases and controls | Variants in genes chosen by hypothesis | Many findings did not replicate |
| GWAS | Thousands of cases and controls | Common variants of small effect | Multiple comparisons; marker variant, not the causal one |
Markers
- GWAS
- SNP
- LOD
- 5 × 10⁻⁸
When to use it
- They have found regions associated with type 2 diabetes, macular degeneration, autoimmune diseases and many cancers.
- Combining many variants gives polygenic risk scores, whose clinical usefulness is still limited.
Common pitfalls
- Using p < 0.05 in a GWAS: with a million tests about 50,000 false positives would appear.
- Comparing cases and controls of different ancestries (population stratification).
- Presenting the associated SNP as the cause of the disease.
- Small samples: the first ORs reported tend to be overestimated (the ‘winner's curse’).
Test yourself
Think of the answer, then tap the question to see it.
Why is the GWAS threshold 5 × 10⁻⁸?
It is the Bonferroni adjustment of 0.05 for about one million independent tests: 0.05 / 1,000,000.
What is population stratification?
Differences in ancestry between cases and controls that create false associations between a SNP and the disease.
Risk allele in 30% of case alleles and 25% of control alleles. Allelic OR?
(0.30/0.70) / (0.25/0.75) = 0.429 / 0.333 ≈ 1.29.
How does linkage differ from association?
Linkage studies joint inheritance of marker and disease within families; association compares frequencies in unrelated cases and controls.
Elsewhere in DurgiPedia
Connections
Educational content on epidemiological and statistical methods; the examples use made-up numbers to teach the calculation. Not a clinical or legal guide: current Colombian rules are in the official sources.
Sources: Genome-Wide Association Studies Fact Sheet (National Human Genome Research Institute, 2020); Genetic Epidemiology (Centers for Disease Control and Prevention, 1997); Bonferroni's method (NIST/SEMATECH e-Handbook of Statistical Methods, National Institute of Standards and Technology, 2012).
Spotted a mistake?We are always improving for the community. Tell us what is wrong and our team will review it.Report a mistake
Representative model, for educational purposes only.