A UNIFIED FRAMEWORK FOR VARIANCE COMPONENT ESTIMATION WITH SUMMARY STATISTICS IN GENOME-WIDE ASSOCIATION STUDIES

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Linear mixed models (LMMs) are crucial for genetic association studies.
• Restricted maximum likelihood estimation (REML) is the standard but has limitations: requires full data, is slow, and yields biased estimates in case-control studies.

Purpose of the Study:

• To develop an alternative framework for variance component estimation in LMMs.
• To address the computational and statistical drawbacks of REML.

Main Methods:

• Introduced MQS (Method of Moments and MINQUE criterion) for variance component estimation.
• Unified Haseman-Elston (HE) regression and LD score regression (LDSC) into a single framework.
• Developed a summary-statistics-based HE form and an exact estimation form of LDSC.

Main Results:

• MQS enables using marginal z-scores from all samples with SNP correlation from a subset, achieving high accuracy.
• The method provides unbiased and statistically efficient estimates.
• Demonstrated computational efficiency for large datasets.

Conclusions:

• MQS offers a computationally efficient, statistically robust alternative to REML for LMMs.
• The method improves SNP heritability estimation and partitioning in population and family studies.
• MQS is implemented in the GEMMA software package.