BIGFAM - variance components analysis from relatives without genotype.

Area of Science:

• Quantitative genetics
• Human genetics
• Statistical genomics

Background:

• Estimating variance components is crucial for understanding complex traits and diseases.
• Current methods often rely on expensive and inaccessible genotype data, limiting their application.
• A genotype-free approach is needed to broaden the scope of variance component analysis.

Purpose of the Study:

• To introduce BIGFAM, a novel genotype-free framework for estimating variance components.
• To assess genetic, shared environmental, and X chromosome effects using only phenotype data from relative pairs.
• To provide a scalable method for variance component analysis in diverse populations.

Main Methods:

• Developed the BIGFAM (Bayesian Inference of Genetic and Family-environment Models) framework.
• Utilized phenotype data from relative pairs within the Generation Scotland and UK Biobank datasets.
• Compared BIGFAM estimates with traditional genotype-based methods.

Main Results:

• BIGFAM estimates showed high correlations with genotype-based methods for heritability (r=0.85) and X chromosome components (r=0.64).
• Identified significant nuclear-family-specific shared environmental effects for dietary-related phenotypes.
• Demonstrated the feasibility of analyzing complex traits without genetic data.

Conclusions:

• BIGFAM offers a powerful and scalable alternative for variance component estimation.
• The framework enables the study of genetic and environmental influences across broader populations.
• This genotype-free approach advances the understanding of complex trait architecture.