GDT

GDT performs family-based genome-wide association testing by leveraging genotype differences among discordant relative pairs to detect disease-associated variants in extended pedigrees.

Key Features:

• Utilization of Discordant Relative Pairs: Incorporates genotype differences from all discordant relative pairs within families, including distant relatives beyond first-degree, to assess genetic associations.
• Robust Statistical Implementation: Employs a robust statistical technique that does not rely on large-sample theory, improving power and accuracy particularly at high significance levels for detecting common and rare disease associations.
• Incorporation of Covariates and Weights: Integrates covariates and weights based on family size into the analysis to account for variation in family structures.
• Extended Pedigree Support: Applicable to extended pedigrees in genome-wide analyses.
• Demonstrated Power Improvements: In computer simulations on extended pedigrees, showed a >13% power improvement in six of eight cases for both common and rare diseases.
• Empirical Validation on T1DGC: Applied to genome-wide linkage data from the Type 1 Diabetes Genetics Consortium (T1DGC) and identified all six strongest associations reported by other studies.
• UBASH3A Association Detection: Detected a Type 1 diabetes association at UBASH3A with p = 4.3 x 10^(-6), stronger than a previously published p = 10^(-4).

Scientific Applications:

• Disease-gene mapping: Family-based genome-wide association studies to map disease-associated variants in extended pedigrees.
• Common and rare variant detection: Identification of both common and rare disease associations at genome-wide significance levels.
• Benchmarking and validation: Comparative evaluation of family-based association methods using computer simulations and empirical datasets such as T1DGC.

Methodology:

Incorporates genotype differences from all discordant relative pairs, employs a robust statistical technique not relying on large-sample theory, and integrates covariates and family-size weights.