ARCHER-LD: Rapid Long-Range Linkage Disequilibrium Calculations at Biobank Scale using GPU Acceleration.

ARCHER-LD computes biobank-scale linkage disequilibrium matrices with over 120 million variants in under 6 hours using GPU acceleration, representing a significant speedup over traditional methods.

• Why it matters: Accurate LD information from individual datasets enhances downstream analyses like fine-mapping, but current computational challenges limit the use of comprehensive, population-specific LD data, especially at biobank scale.
• What they did: The authors developed a multi-GPU distributed computing approach, implemented in Python with CuPy, and applied it to large genomic datasets, achieving up to 28x faster LD calculations compared to existing tools like PLINK.
• The result: This method enables researchers to generate detailed, population-specific LD matrices without external reference panels, facilitating more precise genetic analyses and broadening access to large-scale genomic data.