NetRAX

NetRAX infers maximum-likelihood phylogenetic networks from partitioned multiple sequence alignments to model non-treelike evolution such as hybridization and horizontal gene transfer.

Key Features:

• Efficient Likelihood Computation: Computes the phylogenetic likelihood function on trees and extends these computations to networks using the concept of displayed trees to reduce computational complexity.
• Handling Non-Treelike Evolution: Focuses on inference scenarios that exclude incomplete lineage sorting (ILS), simplifying modeling of reticulate processes.
• Partitioned Multiple Sequence Alignments: Accepts partitioned multiple sequence alignments as input for network inference.
• Output Format: Returns inferred phylogenetic networks in the Extended Newick format.
• Performance and Validation: On simulated data, reports very low relative differences in Bayesian Information Criterion (BIC) versus true networks and near-zero unrooted softwired cluster distance.

Scientific Applications:

• Hybridization inference: Infers network topologies that represent hybridization events in evolutionary histories.
• Horizontal gene transfer detection: Models horizontal gene transfer events across taxa within a network framework.
• Reticulate evolution modeling: Represents other reticulate processes that cannot be captured by bifurcating phylogenetic trees.

Methodology:

Uses maximum-likelihood estimation by computing phylogenetic likelihoods on trees and extending these computations to networks via displayed trees.