MIXSCORE

MIXSCORE integrates case-only admixture signals and case-control SNP association signals to improve detection of disease-associated loci in admixed populations by leveraging admixture-LD and local ancestry-aware imputation.

Key Features:

• Joint association framework: Combines SNP association (LD mapping) with admixture association (mapping by admixture-LD) in a single statistical framework.
• Case-only and case-control signals: Incorporates both case-only admixture signals and case-control SNP association signals for joint analysis.
• Local ancestry-aware imputation: Applies local ancestry-aware imputation for analyses of imputed SNPs.
• Novel scoring statistic: Uses a novel scoring statistic to integrate association and admixture signals.
• Increased statistical power: Empirically increases statistical power by 8% for typed SNPs and 11% for imputed SNPs compared with standard methods.
• Detection of untyped causal variants: Improves detection of disease loci in regions where the causal SNP is untyped and not amenable to conventional imputation.
• Empirical validation: Validated on 6,209 unrelated African Americans from the CARe project genotyped on the Affymetrix 6.0 chip and on a breast cancer GWAS of 5,761 African-American women at the FGFR2 locus.

Scientific Applications:

• Admixed-population GWAS: Enhances genome-wide association studies in admixed populations such as African Americans by jointly modeling admixture-LD and SNP associations.
• Admixture mapping: Performs mapping by admixture-LD to identify loci showing ancestry-associated effects.
• Mapping with imputed data: Improves locus detection using imputed SNPs with local ancestry-aware imputation.
• Mapping where causal SNPs are untyped: Enables identification of disease risk loci in regions lacking typed or conventionally imputable causal variants.

Methodology:

Combines SNP association (LD mapping) with admixture association (mapping by admixture-LD) in a statistical framework that incorporates case-only admixture and case-control signals, applies local ancestry-aware imputation, and computes a novel scoring statistic for typed and imputed SNPs.