The Generalized Higher Criticism for Testing SNP-Set Effects in Genetic Association Studies

Area of Science:

• Genetics
• Biostatistics
• Genomic Epidemiology

Background:

• Complex diseases are influenced by genetic factors, including genes, pathways, and networks.
• Genetic constructs contain multiple single nucleotide polymorphisms (SNPs) that can be correlated and function jointly.
• Identifying the specific subset of SNPs associated with disease risk is challenging due to large numbers and correlations.

Purpose of the Study:

• To propose a novel statistical test, generalized higher criticism (GHC), for assessing the association between SNP sets and disease outcomes.
• To address the limitations of traditional higher criticism tests in genetic association studies, particularly regarding SNP correlations and finite set sizes.

Main Methods:

• Developed the generalized higher criticism (GHC) test to accommodate arbitrary correlation structures among SNPs within a set.
• Enabled accurate analytic p-value calculations for SNP sets of any finite size.
• Derived the detection boundary for the GHC test.

Main Results:

• Empirically compared the power of GHC against existing SNP-set tests using simulations across diverse genetic regions with varying correlation structures and signal sparsity.
• Applied the GHC method to analyze genome-wide association study data for breast cancer from the CGEM study.

Conclusions:

• The GHC test offers an improved approach for detecting SNP-set associations with complex diseases, outperforming existing methods in various scenarios.
• The method is robust to complex correlation structures and finite SNP set sizes commonly found in genetic association studies.