SeaLion

Disentangling signal, conflict and noise

SeaLion is a software framework for evaluating phylogenetic uncertainty in genome-scale datasets. Instead of relying only on support values for a single inferred tree, SeaLion analyzes polarized quartets to measure which relationships are supported, which are contradicted, and where the data remain ambiguous.

The approach was developed to detect misleading signal caused by convergence, plesiomorphy, compositional heterogeneity and branch-length effects. This makes SeaLion especially useful for difficult evolutionary questions in which conventional phylogenomic workflows can produce apparently strong but potentially biased results.

Method development at LIB

At LIB, SeaLion is being developed as a reproducible analysis environment that combines fine-grained quartet evaluation with tree reconstruction and conflict assessment. The project integrates PhyQuart-based signal analysis, the Icebreaker supertree approach, and the RISK and DIST filters to reduce noise and systematically misleading patterns.

The work covers conceptual development, software implementation, workflow optimization, documentation and containerized execution. SeaLion is designed to support transparent, testable and extensible phylogenomic analyses across diverse datasets.

Current applications and next steps

The long-term goal is the steady further development of SeaLion and its polarized quartet framework. Current studies using SeaLion as a core methodological platform address myriapod relationships, phylogenetic conflict in Palaeognathae, the position of Xenacoelomorpha, caecilian phylogeny, and the effects of compositional bias on inference.

Future work will expand scalability, strengthen filtering and diagnostics, improve usability for reproducible research, and apply the framework to additional datasets in biodiversity and evolutionary genomics.