SCOPA

SCOPA applies reverse regression to multivariate genome-wide association studies, using single nucleotide polymorphisms (SNPs) as outcomes and multiple correlated quantitative or categorical phenotypes as predictors to increase power for detecting genetic associations using imputed genotype dosages.

Key Features:

• Multivariate analysis: Performs joint analysis of multiple correlated phenotypes to increase power for detecting SNP associations compared with univariate GWAS.
• Reverse regression GLM: Implements reverse regression within a general linear model where genotype at a specific SNP is the outcome and multiple phenotypes are predictors.
• Phenotype types supported: Handles both quantitative and categorical phenotypes.
• Imputed genotypes (dosage): Accommodates imputed genotypes under a dosage model.
• META-SCOPA integration: Produces association summary statistics that can be integrated by META-SCOPA for meta-analysis across GWAS.

Scientific Applications:

• Multivariate GWAS of correlated traits: Detects and dissects genetic associations across correlated traits, improving discovery beyond univariate approaches.
• Lipid and obesity trait analysis: Applied to high- and low-density lipoprotein cholesterol, triglycerides, and body mass index (BMI) to reveal stronger signals at lipid and obesity loci.
• Meta-analysis across cohorts: Enables consolidation of SCOPA-derived summary statistics via META-SCOPA to increase power across multiple studies.
• Novel locus discovery: Facilitated identification of a genome-wide significant signal at the GPC5 locus for triglycerides (lead SNP rs71427535, p = 1.1x10^-8).

Methodology:

Applies reverse regression implemented as a general linear model with genotype at each SNP as the outcome and multiple phenotypes as predictors, accommodates imputed genotype dosages, and uses META-SCOPA to integrate association summary statistics across studies.