prophaser

prophaser implements probabilistic genotype imputation for ancient DNA (aDNA) using the full Li and Stephens haplotype model to improve genotype recovery from low-coverage and degraded samples.

Key Features:

• Probabilistic Model: Implements the full probabilistic Li and Stephens model of haplotype frequencies to model haplotype structure in fragmented and low-coverage aDNA.
• Improved Accuracy at Low Coverage: Demonstrates improved performance over Beagle-based methods at low coverage and better captures rare genetic variants in ancient samples.
• Massively Parallel Execution: Optimized for parallel processing and efficient execution on GPUs to reduce runtimes for large or computationally intensive imputation tasks.

Scientific Applications:

• Enhancing archaeological genotype resolution: Improves genotype calls from archaeological and other ancient samples with limited or degraded DNA.
• Reconstructing population history: Supports inference of historical population structure and migration patterns from low-coverage aDNA.
• Detecting rare ancient variants: Facilitates identification of rare genetic variants in ancient populations that may relate to past adaptations or diseases.

Methodology:

Evaluations used empirical ancient DNA downsampled to mimic lower coverage scenarios and present-day samples with artificially thinned genotypes.