UNCcombo

UNCcombo performs association testing on next-generation sequencing (NGS) data by integrating genotype likelihood functions (GLF) to bypass genotype calling and account for genotype-calling uncertainty.

Key Features:

• Direct incorporation of genotype likelihoods (GLF): UNCcombo integrates the GLF directly into association analysis to account for genotype-calling uncertainty inherent in NGS data.
• Bypass of genotype calling: The method circumvents intermediate genotype calling to reduce power loss and type-I error inflation associated with genotype-calling uncertainty.
• Likelihood ratio test (LRT) with covariate adjustment: Provides an LRT that supports flexible covariate adjustment to control for confounders in association analyses.
• Enhanced score test: Implements a score test designed to increase statistical power and remain applicable to markers with low minor allele frequency (MAF), including low-frequency variants.
• Combination strategy (UNCcombo): Combines the LRT and score test to balance computational efficiency and power, improving power for quantitative trait analysis by up to approximately 60% for low-frequency causal variants (MAF < 0.01).

Scientific Applications:

• NGS-based association testing: Association testing directly from NGS data without genotype calling, retaining genotype likelihood information in analyses.
• Quantitative trait analysis: Enhances power in quantitative trait association studies, particularly for low-frequency causal variants (MAF < 0.01).
• Common and rare variant analysis: Applicable to both common and rare variant association testing, including studies of complex traits and diseases where low-frequency variants contribute to phenotype.

Methodology:

UNCcombo integrates genotype likelihood functions into likelihood-based association tests (likelihood ratio test and an enhanced score test), combines the two test statistics, and assesses performance via extensive simulations and real data analyses.