Genome-wide association studies (GWAS) are effective in identifying genetic variants associated with disease risk but can be confounded by population stratification, which is systematic ancestry differences between cases and controls. Traditional methods for correcting stratification, such as genomic control and principal components analysis (PCA), are effective when population structure is the only type of sample structure but may fail when family structure or cryptic relatedness is present. Recent advances include methods that account for these complexities, such as family-based association tests and mixed models that incorporate full covariance structures. Population stratification can lead to spurious associations if not properly corrected. Detecting stratification involves computing the Genomic Control λ (λ_GC), which indicates the extent of inflation due to population stratification or other confounders. A λ_GC ≈ 1 indicates no stratification, while λ_GC > 1 suggests stratification or other confounders. PCA and structured association are widely used to correct for stratification, with PCA being computationally tractable for large datasets. However, these methods may not account for family structure or cryptic relatedness, which can lead to inflation in test statistics. Family-based association tests, such as FBAT and QTDT, are immune to stratification because they focus on within-family information. Mixed models, which model population structure, family structure, and cryptic relatedness, have shown superior performance in correcting for these complexities. These models use a mixture of fixed and random effects to account for both heritable and non-heritable variation. Low-frequency and rare variants are increasingly being studied in GWAS, as most genetic heritability remains unexplained. These variants require careful consideration of population structure and relatedness, as deviations from model assumptions can lead to spurious associations. Methods that account for population structure, family structure, and cryptic relatedness are essential for accurate detection of these variants. In conclusion, various methods have been developed to correct for stratification in GWAS, each with its own advantages and limitations. Mixed models, which account for multiple types of sample structure, are particularly effective. For studies with minimal stratification concerns, simple methods without PC covariates may suffice, while more complex methods with PC covariates are recommended for studies with significant stratification. The choice of method depends on the study design and the level of population structure present.Genome-wide association studies (GWAS) are effective in identifying genetic variants associated with disease risk but can be confounded by population stratification, which is systematic ancestry differences between cases and controls. Traditional methods for correcting stratification, such as genomic control and principal components analysis (PCA), are effective when population structure is the only type of sample structure but may fail when family structure or cryptic relatedness is present. Recent advances include methods that account for these complexities, such as family-based association tests and mixed models that incorporate full covariance structures. Population stratification can lead to spurious associations if not properly corrected. Detecting stratification involves computing the Genomic Control λ (λ_GC), which indicates the extent of inflation due to population stratification or other confounders. A λ_GC ≈ 1 indicates no stratification, while λ_GC > 1 suggests stratification or other confounders. PCA and structured association are widely used to correct for stratification, with PCA being computationally tractable for large datasets. However, these methods may not account for family structure or cryptic relatedness, which can lead to inflation in test statistics. Family-based association tests, such as FBAT and QTDT, are immune to stratification because they focus on within-family information. Mixed models, which model population structure, family structure, and cryptic relatedness, have shown superior performance in correcting for these complexities. These models use a mixture of fixed and random effects to account for both heritable and non-heritable variation. Low-frequency and rare variants are increasingly being studied in GWAS, as most genetic heritability remains unexplained. These variants require careful consideration of population structure and relatedness, as deviations from model assumptions can lead to spurious associations. Methods that account for population structure, family structure, and cryptic relatedness are essential for accurate detection of these variants. In conclusion, various methods have been developed to correct for stratification in GWAS, each with its own advantages and limitations. Mixed models, which account for multiple types of sample structure, are particularly effective. For studies with minimal stratification concerns, simple methods without PC covariates may suffice, while more complex methods with PC covariates are recommended for studies with significant stratification. The choice of method depends on the study design and the level of population structure present.