Chaplin

Chaplin performs haplotype inference and tests associations between haplotypes and disease using unphased genotype data from case-control studies.

Key Features:

• Haplotype inference methods: Implements the prospective score test of Schaid et al., a modification of the prospective estimating equations by Zhao et al., and the retrospective likelihood method of Epstein and Satten for haplotype inference from unphased genotype data in case-control studies.
• Bias and power comparison: Compares the three methods in terms of bias and statistical power when applied to unphased genotype data from case-control studies.
• Model-specific performance: Shows comparable performance among methods under a multiplicative haplotype effect model, with the retrospective likelihood method demonstrating increased efficiency under dominant and recessive models.
• Hardy-Weinberg Equilibrium considerations: Assumes Hardy-Weinberg Equilibrium (HWE) for haplotype frequencies and reports that prospective methods are robust to HWE deviations while the retrospective likelihood method can be biased under dominant and recessive models when HWE is violated.
• Fixation index adjustment: Incorporates a modification that introduces a non-negative fixation index applicable across all haplotype pairs to reduce bias of the retrospective likelihood method when HWE does not hold.

Scientific Applications:

• Genetic association studies: Enables haplotype-based association testing in case-control genetic epidemiology studies.
• Identifying genetic risk factors: Identifies specific haplotypes or haplotype features associated with disease risk from genotype data.
• Elucidating disease pathways: Supports analyses that link haplotype associations to disease mechanisms.
• Informing personalized medicine: Provides haplotype association information that can inform personalized medicine strategies.

Methodology:

Uses the prospective score test of Schaid et al., a modification of the prospective estimating equations by Zhao et al., and the retrospective likelihood method of Epstein and Satten; compares these methods for bias and power on unphased genotype data from case-control studies and implements a non-negative fixation index to adjust the retrospective likelihood when Hardy-Weinberg Equilibrium is violated.