GWAS

A genome-wide association study (GWAS) scans many genetic variants across the genome for statistical association with a trait. Related tools in this library include polygenic scores built from GWAS weights and Mendelian randomisation that uses genetic instruments (often from GWAS) to probe putative causal links under explicit assumptions.

GWAS hits, polygenic scores, and MR estimates are easy to collapse into “genes cause the trait.” They answer different questions—discovery of associations, prediction accuracy, and instrument-based causal inference—with different failure modes.

Claim ledger

What the evidence shows

Drawn from 7 studies in this library. Mix labels say which citation roles are present; they are not a strength score. Supports means evidence for a finding; Challenges means evidence against a stated position; Qualifies marks scope.

• ○ Limited citations

  Polygenic score accuracy can differ a lot among groups that share a broad ancestry label.

  Polygenic score accuracy can differ substantially among groups that share a broad ancestry label; performance tracks ascertainment and SES-related structure, not ancestry labels alone.

  1 supporting · 2 qualifying

  Supports

  1. 1Do polygenic scores work equally within one ancestry?

  Qualifies

  1. 1Holstein GWAS across 31 dairy traits— different species and trait system — dairy cattle GWAS, not human PGS accuracy
  2. 2Mendelian randomisation associates specific gut taxa with several cancers— different question — MR disease associations, not PGS prediction accuracy

  Study comparison

  Study	Role	Design	N	Population	Outcome
  Do polygenic scores work equally within one ancestry?2020	Supports	Computational / modellingWithin-ancestry stratification of PGS accuracy in UK Biobank (education, height, BMI, etc.)	N=408434 · 408,434 UK Biobank participants passing QC; analyses often within White British strata	UK Biobank participants of broadly similar genetic ancestry	Within-ancestry variation in polygenic score prediction accuracy
  Holstein GWAS across 31 dairy traits2011	Qualifiesdifferent species and trait system — dairy cattle GWAS, not human PGS accuracy	Computational / modellingGWAS of ~46k SNPs against PTA traits in contemporary US Holsteins	N=1654 · 1,654 Holstein cows; 45,878 genotyped SNPs	Contemporary US Holstein cattle with predicted transmitting abilities	Additive SNP associations with production, health, and reproduction traits
  Mendelian randomisation associates specific gut taxa with several cancers2023	Qualifiesdifferent question — MR disease associations, not PGS prediction accuracy	Mendelian randomisationTwo-sample MR of MiBioGen gut taxa instruments against eight cancer GWAS summary sets	211 taxa instruments; cancer GWAS sample sizes vary by malignancy — no single primary N	Gut microbiome and cancer GWAS summary statistics (IEU OpenGWAS/consortia)	Causal microbiome–cancer associations (stringent and sensitivity analyses)

• ○ Limited citations

  Liver eQTL maps link common variants to expression of disease-relevant metabolic genes.

  Human liver eQTL mapping links common variants to expression of disease-relevant metabolic pathway genes, showing one route from GWAS-scale variation to molecular intermediates.

  1 supporting · 2 qualifying

  Supports

  1. 1Genetics of gene expression in human liver

  Qualifies

  1. 1Can rare variants fake common GWAS hits?— different question — how rare variants can create synthetic common signals
  2. 2Mendelian randomisation associates specific gut taxa with several cancers— different endpoint — cancer odds via MR, not liver expression QTLs

  Study comparison

  Study	Role	Design	N	Population	Outcome
  Genetics of gene expression in human liver2008	Supports	Cross-sectionalLiver expression and genotype profiling to map hepatic eQTLs and coexpression networks	N=427 · Human liver cohort of 427 Caucasian subjects	Human liver tissue donors	Genetic architecture of hepatic gene expression (eQTLs/networks)
  Can rare variants fake common GWAS hits?2010	Qualifiesdifferent question — how rare variants can create synthetic common signals	Computational / modellingGenealogical simulations of rare causal variants generating synthetic common-SNP GWAS signals	Simulation study (~30% of runs detect genome-wide associations) — no empirical sample N	Simulated genealogies with rare disease alleles	Synthetic genome-wide associations created by rare causal variants
  Mendelian randomisation associates specific gut taxa with several cancers2023	Qualifiesdifferent endpoint — cancer odds via MR, not liver expression QTLs	Mendelian randomisationTwo-sample MR of MiBioGen gut taxa instruments against eight cancer GWAS summary sets	211 taxa instruments; cancer GWAS sample sizes vary by malignancy — no single primary N	Gut microbiome and cancer GWAS summary statistics (IEU OpenGWAS/consortia)	Causal microbiome–cancer associations (stringent and sensitivity analyses)

• ○ Limited citations

  Rare variants can create synthetic GWAS signals at common markers when clustered.

  Rare variants can collectively create synthetic genome-wide association signals at common markers when they are not extremely numerous and evenly spread—so a GWAS peak is not automatically a single common causal allele.

  1 supporting · 2 qualifying

  Supports

  1. 1Can rare variants fake common GWAS hits?

  Qualifies

  1. 1Holstein GWAS across 31 dairy traits— empirical cattle GWAS peaks — not a synthetic-signal simulation
  2. 2Do polygenic scores work equally within one ancestry?— different question — PGS transportability, not signal synthesis

  Study comparison

  Study	Role	Design	N	Population	Outcome
  Can rare variants fake common GWAS hits?2010	Supports	Computational / modellingGenealogical simulations of rare causal variants generating synthetic common-SNP GWAS signals	Simulation study (~30% of runs detect genome-wide associations) — no empirical sample N	Simulated genealogies with rare disease alleles	Synthetic genome-wide associations created by rare causal variants
  Holstein GWAS across 31 dairy traits2011	Qualifiesempirical cattle GWAS peaks — not a synthetic-signal simulation	Computational / modellingGWAS of ~46k SNPs against PTA traits in contemporary US Holsteins	N=1654 · 1,654 Holstein cows; 45,878 genotyped SNPs	Contemporary US Holstein cattle with predicted transmitting abilities	Additive SNP associations with production, health, and reproduction traits
  Do polygenic scores work equally within one ancestry?2020	Qualifiesdifferent question — PGS transportability, not signal synthesis	Computational / modellingWithin-ancestry stratification of PGS accuracy in UK Biobank (education, height, BMI, etc.)	N=408434 · 408,434 UK Biobank participants passing QC; analyses often within White British strata	UK Biobank participants of broadly similar genetic ancestry	Within-ancestry variation in polygenic score prediction accuracy

• ○ Limited citations

  Large dairy GWAS recover many additive SNP effects, some in genes and some elsewhere.

  Large GWAS of Holstein dairy traits recovered thousands of genome-wide significant additive effects, with many top SNPs in genes and some SNPs associated with multiple traits (pleiotropy).

  1 supporting · 2 qualifying

  Supports

  1. 1Holstein GWAS across 31 dairy traits

  Qualifies

  1. 1Soybean GWAS networks for agronomy— different species — soybean agronomic networks, not dairy PTA traits
  2. 2Soybean GWAS for protein and oil— different traits — soybean protein/oil, not cattle production health

  Study comparison

  Study	Role	Design	N	Population	Outcome
  Holstein GWAS across 31 dairy traits2011	Supports	Computational / modellingGWAS of ~46k SNPs against PTA traits in contemporary US Holsteins	N=1654 · 1,654 Holstein cows; 45,878 genotyped SNPs	Contemporary US Holstein cattle with predicted transmitting abilities	Additive SNP associations with production, health, and reproduction traits
  Soybean GWAS networks for agronomy2017	Qualifiesdifferent species — soybean agronomic networks, not dairy PTA traits	Computational / modellingGWAS of diverse soybean landraces/cultivars phenotyped across locations and years	N=809 · 809 accessions; >10 million SNPs/indels after imputation	Diverse Glycine max landraces and cultivars	Genetic networks underlying agronomical traits including flowering and yield components
  Soybean GWAS for protein and oil2014	Qualifiesdifferent traits — soybean protein/oil, not cattle production health	Computational / modellingGWAS of soybean germplasm SNPs for seed protein and oil with LD/structure estimation	N=298 · 298 germplasm accessions; 31,954 QC SNPs	Soybean germplasm accessions	SNP associations with seed protein and oil content

• ○ Limited citations

  Mendelian randomisation on gut taxa reports a limited set of microbiome–cancer links.

  Using genetic instruments for gut microbiota taxa, Mendelian randomisation reported eleven stringent microbiome-to-cancer associations (directions sometimes opposing across cancers for related taxa)—an MR application that depends on GWAS-quality instruments and their assumptions.

  1 supporting · 2 qualifying

  Supports

  1. 1Mendelian randomisation associates specific gut taxa with several cancers

  Qualifies

  1. 1Do polygenic scores work equally within one ancestry?— different question — PGS accuracy within ancestry, not MR cancer odds
  2. 2Genetics of gene expression in human liver— different intermediate — liver eQTLs, not gut-taxa instruments

  Study comparison

  Study	Role	Design	N	Population	Outcome
  Mendelian randomisation associates specific gut taxa with several cancers2023	Supports	Mendelian randomisationTwo-sample MR of MiBioGen gut taxa instruments against eight cancer GWAS summary sets	211 taxa instruments; cancer GWAS sample sizes vary by malignancy — no single primary N	Gut microbiome and cancer GWAS summary statistics (IEU OpenGWAS/consortia)	Causal microbiome–cancer associations (stringent and sensitivity analyses)
  Do polygenic scores work equally within one ancestry?2020	Qualifiesdifferent question — PGS accuracy within ancestry, not MR cancer odds	Computational / modellingWithin-ancestry stratification of PGS accuracy in UK Biobank (education, height, BMI, etc.)	N=408434 · 408,434 UK Biobank participants passing QC; analyses often within White British strata	UK Biobank participants of broadly similar genetic ancestry	Within-ancestry variation in polygenic score prediction accuracy
  Genetics of gene expression in human liver2008	Qualifiesdifferent intermediate — liver eQTLs, not gut-taxa instruments	Cross-sectionalLiver expression and genotype profiling to map hepatic eQTLs and coexpression networks	N=427 · Human liver cohort of 427 Caucasian subjects	Human liver tissue donors	Genetic architecture of hepatic gene expression (eQTLs/networks)

Debates

Tensions and limits

Some items are genuine disagreements on the same question. Others mark different assays, populations, or outcomes.

Timeline

How understanding moved

Study years are when the paper was published. Evidence edits are dated changes to this page's claims. Explanations are when PaperFren added a Discovery — not a claim that the science happened that day.

Papers

7 studies in this library bear on GWAS, ordered by citations.

• Genetics of gene expression in human liver

  A 427-person liver cohort maps eQTLs and expression networks that connect genetic variation to metabolic disease biology.

  PLoS biology · 2008 · 759 citations

• Can rare variants fake common GWAS hits?

  Collections of rare causal variants can create synthetic genome-wide association signals at common SNPs.

  PLoS biology · 2010 · 675 citations

• Do polygenic scores work equally within one ancestry?

  Even within a relatively homogeneous ancestry group, PGS prediction accuracy differs substantially across socio-economic and related strata.

  eLife · 2020 · 400 citations

• Mendelian randomisation associates specific gut taxa with several cancers

  Genetic instruments for gut microbiota taxa showed causal associations with breast, lung, colorectal, prostate, gastric, and head/neck cancers—sometimes in opposing directions for the same genus.

  BMC medicine · 2023 · 346 citations

• Holstein GWAS across 31 dairy traits

  Genome-wide SNP association in U.S. Holsteins links many loci to production, health, fertility, and type traits.

  BMC genomics · 2011 · 336 citations

• Soybean GWAS networks for agronomy

  Large-scale soybean resequencing GWAS maps genetic networks underlying major agronomic traits.

  Genome biology · 2017 · 312 citations

• Soybean GWAS for protein and oil

  Genome-wide SNPs in 298 soybean accessions associate with seed protein and oil content.

  BMC genomics · 2014 · 308 citations

Questions

What is still open

1. Human PGS accuracy, liver eQTLs, rare-variant synthetic-signal theory, livestock/plant GWAS, and microbiome–cancer MR all live under a GWAS umbrella, but they answer different estimands. Pooling them as one “GWAS finding” erases species, trait, and method limits.

   Watch this

2. A genome-wide significant hit is not the same claim as a portable polygenic score or an MR causal estimate. Synthetic-signal work limits naive readings of peaks; PGS work limits naive portability; MR adds instrument assumptions on top of association.

   Watch this

Ask PaperFren about GWAS