High-definition likelihood inference of genetic correlations across human complex traits

Area of Science:

• Genetics
• Bioinformatics
• Statistical Genomics

Background:

• Understanding shared genetic architecture between complex traits is crucial.
• Genome-wide association studies (GWAS) provide summary statistics for genetic analyses.
• Linkage Disequilibrium Score Regression (LDSC) is a common method for estimating genetic correlations but has limitations.

Purpose of the Study:

• To develop a novel method, High-Definition Likelihood (HDL), for more precise estimation of genetic correlations.
• To improve upon existing methods by fully utilizing Linkage Disequilibrium (LD) information across the genome.

Main Methods:

• Developed the High-Definition Likelihood (HDL) method for genetic correlation estimation.
• Utilized summary statistics from genome-wide association studies (GWAS).
• Applied HDL and LDSC to estimate 435 genetic correlations among 30 phenotypes in the UK Biobank (UKBB).

Main Results:

• HDL reduces the variance of genetic correlation estimates by approximately 60% compared to LDSC.
• This variance reduction is equivalent to a 2.5-fold increase in sample size.
• HDL identified 57 additional significant genetic correlations not detected by LDSC, significantly outperforming LDSC in discovering trait associations.

Conclusions:

• HDL provides a more powerful and precise approach for estimating genetic correlations.
• The method enhances the ability to uncover shared genetic architecture among human complex traits.
• HDL offers a significant advancement for genomic analyses and understanding trait relationships.