HAPI-UR

HAPI-UR infers phased haplotypes from unphased genotypes to enable population-scale human genetics analyses.

Key Features:

• Scalability and Efficiency: Capable of processing datasets with over 100,000 samples and demonstrates improved computational efficiency relative to Beagle, MaCH, IMPUTE2, and SHAPEIT.
• Accuracy and Error Reduction: Achieves high phasing accuracy with lower switch-error rates than most competitors (except Beagle); consensus phasing by running HAPI-UR three times can reduce switch-error rate below Beagle.
• Versatility: Supports phasing of unrelated samples and family data, including trios and duos.
• Performance Across Diverse Data Sets: Tested on datasets comprising multiple ethnicities and shows consistent switch-error reduction as sample size increases across diverse populations.
• Runtime Efficiency: Runs up to 18 times faster than other methods and exhibits superior runtime scaling to Beagle, enabling practical application at higher marker densities.

Scientific Applications:

• Genome-Wide Association Studies (GWAS): Provides accurate phased haplotypes to improve power for detecting associations between genetic variants and traits.
• Population Genetics: Enables analysis of population structure and history through detailed haplotype information across diverse populations.
• Personalized Medicine: Facilitates identification of disease-associated variants in diverse populations to support development of tailored therapeutic strategies.

Methodology:

Uses a novel algorithmic phasing approach that optimizes the phasing process to reduce switch errors and supports consensus phasing by running HAPI-UR three times.