ACG

ACG infers ancestral recombination graphs and population genetic parameters from aligned nucleotide sequence data to incorporate recombination into evolutionary inference.

Key Features:

• Ancestral Recombination Graph (ARG): Models evolutionary history by explicitly representing recombination events rather than a single phylogenetic tree.
• Bayesian Markov Chain Monte Carlo (MCMC): Uses Bayesian MCMC to estimate posterior likelihoods of evolutionary models from genetic alignments.
• Full Felsenstein Likelihood Computation: Computes the full Felsenstein likelihood of the ARG instead of relying on pairwise or composite likelihoods.
• Computational Efficiency: Implements computational strategies to accelerate inference and is reported to be approximately 100 times faster than comparable recombination-aware programs.
• Estimation of Population Parameters: Estimates posterior distributions of scaled population size and recombination rate and reconstructs recombinant history including recombination breakpoints, TMRCA distributions, and non-recombining trees at specific sites.
• Model Flexibility: Supports multiple nucleotide substitution models and multiple population size models.

Scientific Applications:

• Population History Reconstruction: Reconstructs historical demography and evolutionary processes while accounting for recombination.
• Genetic Diversity Studies: Quantifies how recombination influences genetic diversity within and between populations using sequence alignments.
• Evolutionary Biology Research: Enables investigation of molecular evolutionary mechanisms in organisms with complex genomic architectures by integrating recombination into inference.

Methodology:

Implements ARG modeling with Bayesian MCMC to sample posterior distributions from aligned nucleotide data, computes the full Felsenstein likelihood on the ARG, estimates posterior distributions of scaled population size and recombination rate, infers recombination breakpoints, TMRCA distributions, and non-recombining site-specific trees, and supports multiple substitution and population size models while applying computational optimizations for speed.