EagleImp

EagleImp performs genome-wide phasing and genotype imputation to enable accurate and efficient imputation from large reference panels.

Key Features:

• Accelerated Performance: Achieves runtimes 2 to 10 times faster than Eagle2 and PBWT depending on single- or multiprocessor configuration.
• High Accuracy: Maintains or surpasses the phasing and imputation quality of preceding methods when using extensive reference datasets, including panels with over one million genomes.
• Advanced Technical Optimizations: Implements automated chromosome splitting, dynamic memory management, and fast reading/writing of large files, and provides user-configurable algorithm and output options.
• Sublinear Runtime Behavior: Employs algorithmic optimizations that exhibit sublinear theoretical runtime behavior for scaling to whole-genome datasets.
• Reference-based Phasing and Imputation: Performs reference-based phasing and genotype imputation on large-scale genomic data.

Scientific Applications:

• Genome-wide Association Studies (GWAS): Imputes common variants for GWAS with accuracy comparable to or exceeding the Sanger Imputation Service, Michigan Imputation Server, and TOPMed Imputation Server.
• Large Reference Panels: Scales to analyses involving very large reference panels, accommodating more than one million genomes.
• Reference Panel Processing: Applicable to whole-genome sequencing reference panels such as the 1000 Genomes Project and the Haplotype Reference Consortium.

Methodology:

Performs reference-based phasing and genotype imputation using algorithmic optimizations that yield sublinear theoretical runtime behavior, combined with automated chromosome splitting, dynamic memory management, and fast I/O for large files.