StocSum: stochastic summary statistics for whole genome sequencing studies

Area of Science:

• Genomics
• Statistical Genetics
• Bioinformatics

Background:

• Genomic summary statistics are crucial for genome-wide association studies (GWAS) but often require external reference panels for linkage disequilibrium (LD) information.
• Existing reference panels struggle to accurately represent LD structures in underrepresented, admixed populations, and for rare variants from whole genome sequencing (WGS) studies.
• This limitation restricts the application scope of genomic summary statistics, hindering genetic research in diverse populations.

Approach:

• Introducing StocSum, a novel reference-panel-free statistical framework for generating, managing, and analyzing stochastic summary statistics.
• Utilizes random vectors to capture necessary correlation information without external panels.
• Enables a wide range of downstream applications directly from summary statistics.

Key Points:

• StocSum supports diverse applications including single-variant tests, conditional association tests, gene-environment interaction tests, variant set tests, meta-analysis, and LD score regression.
• Demonstrates accuracy and computational efficiency using cohorts from the Trans-Omics for Precision Medicine Program.
• Overcomes the limitations of external reference panels for LD structure, especially for WGS data.

Conclusions:

• StocSum provides a robust, reference-panel-free solution for analyzing genomic summary statistics.
• Facilitates broader sharing and utilization of WGS-derived summary statistics, particularly for underrepresented and admixed populations.
• Advances precision medicine by enabling more inclusive and comprehensive genetic analyses.