STELLS

STELLS infers population trees representing past population divergence histories using haplotype-based coalescent likelihood methods.

Key Features:

• STELLSH (Species Tree Estimation using Likelihood of Linked Haplotypes): implements a coalescent likelihood approach that leverages linked haplotypes rather than treating variants as unlinked.
• Haplotype-based analysis: uses haplotypes over multiple single nucleotide polymorphisms (SNPs) within non-recombining regions to retain linkage information.
• Approximate likelihood model: employs an approximated likelihood to improve computational efficiency and enable scaling to large and whole-genome datasets.
• Non-Monte Carlo computation: computes likelihoods without relying on Monte Carlo methods to reduce computational cost.
• Validation on real and simulated data: performance has been evaluated using simulation data and the 1000 Genomes Project dataset demonstrating accuracy and efficiency.

Scientific Applications:

• Population divergence inference: reconstructs population trees to infer historical divergence events among populations.
• Population evolutionary studies: supports analyses of evolutionary processes by preserving linkage information across SNPs.
• Large-scale genomic analyses: applicable to large population genetic projects and whole-genome datasets, including the 1000 Genomes Project.

Methodology:

STELLS uses the STELLSH coalescent likelihood framework applied to haplotypes across multiple SNPs in non-recombining regions, implements an approximated likelihood model for scalability, and avoids Monte Carlo likelihood computation.