domains / families / science / association-testing

association-testing

separate

Single-variant common-variant GWAS with plink2 --glm (linear/logistic, Firth) and the linear mixed models GEMMA, BOLT-LMM, SAIGE, regenie (SPA). A GWAS statistic is valid only when genotype is independent of unmodeled phenotype drivers after the chosen covariates and random effects, so the engine follows sample structure and case:control imbalance, not taste: PC covariates absorb continuous ancestry but cannot remove relatedness (a covariance structure needing an LMM), genomic inflation above 1 is mostly true polygenic signal not confounding (the LDSC intercept is the diagnostic), LOCO prevents proximal contamination, and SPA/Firth keep the tail calibrated at extreme imbalance and low MAC. Use when running single-variant GWAS, choosing between a GLM and a mixed model, or controlling stratification, relatedness, and case:control imbalance. For rare-variant aggregation (burden, SKAT, SKAT-O, ACAT) see rare-variant-association; fine-mapping and MR see causal-genomics; PRS see clinical-da…

When to use

• single-variant GWAS
• choosing between GLM and mixed model
• controlling stratification
• controlling relatedness
• controlling case:control imbalance
• linear mixed models
• SPA and Firth correction

When not to use

• rare-variant aggregation
• burden tests
• SKAT
• fine-mapping
• Mendelian randomization
• polygenic risk scores

Skill page

Source