Leveraging the local genetic structure for trans-ancestry association mapping

Area of Science:

• Genetics
• Statistical Genetics
• Population Genetics

Background:

• Genome-wide association studies (GWASs) have significantly advanced complex trait genetics but suffer from a lack of diversity, primarily using European ancestry samples.
• This ancestry bias limits the generalizability of findings and exacerbates health disparities.
• Trans-ancestry association mapping (TRAM) presents a strategy to integrate diverse population data, but requires robust statistical methods.

Purpose of the Study:

• To introduce LOG-TRAM, a novel statistical method designed to improve trans-ancestry association mapping (TRAM) by utilizing local genetic architecture.
• To enhance the power of identifying genetic risk variants in under-represented populations.
• To enable more accurate polygenic risk score construction for diverse ancestries.

Main Methods:

• Development and application of LOG-TRAM, a statistical framework for TRAM that leverages local genetic architecture.
• Utilizing large-scale biobank datasets, including BioBank Japan, UK Biobank, and African population data.
• Evaluating LOG-TRAM's performance in terms of statistical power, p-value calibration, and correction of confounding biases.

Main Results:

• LOG-TRAM significantly improves statistical power for detecting risk variants in under-represented populations.
• The method produces well-calibrated p-values, ensuring reliable association findings.
• Application of LOG-TRAM to diverse GWAS summary statistics demonstrated substantial power gains and effective bias correction.
• LOG-TRAM successfully identified ancestry-specific loci and facilitated the creation of more accurate polygenic risk scores.

Conclusions:

• LOG-TRAM is an effective statistical method for advancing trans-ancestry association mapping.
• The approach enhances the discovery of genetic variants across diverse populations, addressing critical gaps in current GWAS.
• LOG-TRAM holds significant potential for reducing health disparities by enabling more equitable genetic research and risk prediction.