CollapsABEL

CollapsABEL implements the Generalized Compound Double Heterozygosity (GCDH) test to detect compound heterozygosity effects in genome-wide association studies (GWAS) for binary and continuous traits with covariates and to help address missing heritability.

Key Features:

• Generalized Linear Models: Uses generalized linear models to analyze both binary and continuous traits while adjusting for covariates.
• GCDH test: Implements the Generalized Compound Double Heterozygosity test to extend traditional approaches for detecting compound heterozygosity effects.
• High computational efficiency: Core computations are optimized via implementations in Java or C++ to handle large-scale GWAS datasets.
• Data management: Stores intermediate results using SQL databases or big.matrix files to reduce memory footprint.
• R-based analytical components: Provides genotype collapsing, statistical testing, power estimation, type I error control, and graphics generation implemented in R.
• Enhanced detection power: Simulation studies demonstrate superior power to detect CH-like interactions compared with conventional single-SNP approaches.
• Application to complex phenotypes: Improves detection of CH alleles for investigations of complex human phenotypes and missing heritability.

Scientific Applications:

• GWAS interaction discovery: Detects compound heterozygosity and CH-like effects contributing to complex trait variation in genome-wide association studies.
• Missing heritability studies: Identifies allelic configurations that may explain phenotypic variance not captured by single-variant analyses.
• Genetic architecture of complex phenotypes: Supports analyses aimed at elucidating multi-allelic contributions to human phenotypes.

Methodology:

Implements the GCDH test using generalized linear models; core routines are optimized in Java or C++; intermediate results are stored in SQL databases or big.matrix files; genotype collapsing, statistical testing, power estimation, type I error control, and graphics generation are provided in R.