LSX

LSX reduces lineage rate heterogeneity in multi-gene phylogeny inference by implementing LS³ and LS⁴ subsampling algorithms that remove fast- and extremely slow-evolving sequences to improve phylogenetic signal.

Key Features:

• LS³ algorithm: Identifies gene-specific subsets by removing fast-evolving sequences to obtain relatively homogeneous evolutionary rates, minimizing non-phylogenetic noise while preserving phylogenetic signal.
• LS⁴ algorithm: Detects both fast-evolving and extremely slow-evolving sequences to enable less stringent subsampling in datasets containing very slow-evolving lineages.
• Data subsampling strategy: Performs sequence removal/subselection across genes to reduce lineage rate heterogeneity in multi-gene datasets.
• Reprogrammed lineage rate calculations: Revises LS³ implementation to improve lineage rate calculations compared with the previous implementation.
• R implementation: Implemented as an R script.
• Scalability: Applicable to multi-gene sequence datasets of any size.

Scientific Applications:

• Multi-gene phylogeny inference: Reduces bias from lineage rate heterogeneity to improve accuracy of phylogenetic tree estimation.
• Resolution of contentious nodes: Enhances resolution of difficult or disputed phylogenetic relationships by minimizing non-phylogenetic signal.
• Empirical phylogenetic analyses: Applied to a catfish phylogeny dataset exhibiting high lineage rate heterogeneity to improve inference.

Methodology:

Implements LS³ and LS⁴ algorithms for gene-specific sequence subsampling (LS³ removes fast-evolving sequences; LS⁴ detects fast- and extremely slow-evolving sequences) and its performance has been validated via simulations and empirical studies.