SCORE-Seq

SCORE-Seq detects associations between rare genetic variants and disease phenotypes in high-throughput sequencing studies by aggregating mutation information across variant sites and performing score-based association tests.

Key Features:

• Aggregation of Mutation Information: Aggregates mutation information across multiple variant sites within a gene using a weighted linear combination.
• Flexible Regression Models: Fits regression models relating aggregated genetic scores to phenotypes for case-control, cross-sectional, cohort, and family study designs and supports binary, quantitative, and age-at-onset outcomes.
• Incorporation of Covariates: Includes arbitrary covariates, including environmental factors and ancestry variables, in regression models.
• Theoretically Optimal Procedures: Derives theoretically optimal procedures for combining rare mutations and constructs test statistics to maximize statistical power and computational efficiency.
• Flexible Allele-Frequency Thresholds: Supports fixed and variable allele-frequency thresholds for defining rare variants.
• Directional Effects Analysis: Handles combined rare-variant effects that are in the same or opposite directions.
• Statistical Significance Assessment: Uses score-type statistics with significance assessed via asymptotic normal approximation or resampling methods.

Scientific Applications:

• Complex Disease Genetics: Tests gene-level rare variant associations to uncover genetic contributions to complex human diseases.
• Cross-Study and Phenotype Analyses: Applies to diverse study designs and phenotypic analyses, including case-control, cohort, family studies, binary outcomes, quantitative traits, and age-at-onset phenotypes.
• Drug-Target Deep Resequencing: Has been applied to deep-resequencing of drug targets to identify rare variants associated with total cholesterol levels.

Methodology:

Aggregates mutation information across multiple sites within a gene using a weighted linear combination, fits regression models relating aggregated genetic scores to phenotypes with inclusion of covariates and allele-frequency thresholds, and assesses significance using score-type statistics with asymptotic normal approximation or resampling methods.