Multiple linear combination (MLC) regression tests for common variants adapted to linkage disequilibrium structure.

Area of Science:

• Statistical Genetics
• Genomic Association Studies

Background:

• Traditional genetic association studies analyze variants individually, potentially limiting power and interpretability.
• Joint analysis of multiple variants within a gene, using gene-based methods, can enhance statistical power and robustness.

Purpose of the Study:

• To investigate the performance of multiple linear combination (MLC) test statistics for analyzing common variants in the presence of linkage disequilibrium (LD).
• To compare MLC with existing methods like minimum P-value, variance-component, and principal-component approaches.

Main Methods:

• Developed MLC test statistics that exploit gene-specific LD structure by clustering correlated variants.
• Recoded variants within clusters to ensure positive pairwise correlations and combined effects linearly.
• Aggregated cluster effects using a quadratic sum of squares and cross-products, yielding a test statistic with reduced degrees of freedom.
• Conducted simulation studies using 1000 genes from HapMap Asian haplotypes under realistic trait models.

Main Results:

• MLC demonstrated well-powered and robust performance across diverse gene structures compared to existing methods.
• MLC's mean power was comparable or superior to other methods, particularly when multiple causal variants were present.
• MLC exhibited less variation in gene-specific test size and power across the 1000 simulated genes, indicating greater consistency.
• The cluster construction facilitated interpretation of within-gene LD structure as haplotypic effects.

Conclusions:

• MLC is a powerful and robust gene-based statistical method for genetic association analysis of common variants.
• MLC offers a complementary approach for genome-wide discovery, showing consistent performance and aiding in the interpretation of complex genetic architectures.
• The method effectively leverages LD structure to improve the analysis and interpretation of genetic associations within genes.