Association score testing for rare variants and binary traits in family data with shared controls

Area of Science:

• Genetics
• Statistical genetics
• Genomics

Background:

• Genome-wide association studies (GWAS) primarily focus on common variants to identify trait loci.
• The rare variant-common disease hypothesis suggests rare variants contribute to missing heritability.
• Advances in sequencing and statistical methods enable large-scale rare variant analysis.

Purpose of the Study:

• To extend existing score tests (e.g., χcorrected2, WQLS, SKAT) to a multiple variant association framework.
• To evaluate and compare the performance of these extended score tests against linear mixed models (LMMs).
• To demonstrate the flexibility and applicability of these tests in various study designs, including those using publicly available allele frequencies.

Main Methods:

• Extension of score tests (χcorrected2, WQLS, SKAT) for multiple rare variant association.
• Comparison of statistical performance against linear mixed models (LMMs).
• Formulation of tests as differences in estimated marker allele frequencies between affected and unaffected groups.

Main Results:

• Extended score tests demonstrate statistical performance comparable to LMMs.
• The proposed methods are flexible, accommodating related, unrelated, or mixed subject groups.
• Feasibility of designs using subsets of affected subjects and public control allele frequencies is shown.
• Significant impact of linkage disequilibrium on test performance is identified.

Conclusions:

• Extended score tests offer a flexible and powerful alternative to LMMs for rare variant association studies.
• These methods facilitate novel study designs, enhancing the ability to uncover genetic contributions to complex diseases.
• Understanding linkage disequilibrium is crucial for optimizing the performance of these rare variant association tests.