Ohana

Ohana infers individual ancestry components and estimates population trees from genotype data using a maximum likelihood optimization of the STRUCTURE model while handling genotype likelihoods from NGS data.

Key Features:

• Optimization Algorithm for STRUCTURE Model: Implements a novel optimization algorithm tailored to the classical STRUCTURE model within a maximum likelihood framework, achieving higher likelihood solutions with comparable computational efficiency.
• Genotype Likelihoods Integration: Incorporates genotype likelihoods to account for genotype-calling uncertainty associated with next-generation sequencing (NGS) data.
• Population Tree Estimation: Employs a Gaussian approximation to estimate population trees from inferred ancestry components and visualize evolutionary relationships among populations.
• Simulation-Based Validation: Validated using coalescence simulations of diverging populations to assess accuracy of ancestry component inference and population tree topologies.
• Visualization Tools: Provides visualization for interpreting inferred ancestry components and population tree structures.

Scientific Applications:

• Fractional Ancestry Classification: Assigns individuals fractionally into discrete ancestry components.
• Admixture Inference: Infers historical admixture events and estimates their timelines.
• Phylogenetic Reconstruction: Constructs population trees reflecting genetic divergence among populations.

Methodology:

Optimization of the STRUCTURE model within a maximum likelihood framework, handling of genotype likelihoods from NGS data, population tree estimation via a Gaussian approximation, and validation through coalescence simulations.