In a new study involving whole-genome data, researchers present "CASTER," a tool that uses arrangements in DNA sequences known as site patterns to infer "species trees," which are diagrams that depict the evolutionary relationships among species. The tool, which performs with exceptional accuracy and scalability and overcomes the limitations of traditional phylogenetic methods, offers transformative potential for evolutionary research. The growing availability of genomic data has revitalized efforts to construct precise species trees and model gene tree variations. However, the methodology for utilizing genome-wide data lags behind data availability. While traditional methods struggle with incomplete lineage sorting (ILS), and two-step approaches face computational hurdles, emerging site-based methods address ILS but are hampered by scalability and accuracy issues. In response, Chao Zhang and colleagues developed CASTER (Coalescence-aware Alignment-based Species Tree Estimator), a novel approach that directly infers species trees from whole-genome alignments. Through extensive simulations and analyses of previously studied genomic datasets, including for birds and mammals, Zhang et al. demonstrate that CASTER is typically faster and more accurate than other state-of-the-art methods in analyzing hundreds of recombining genomes with precise phylogenic inference. However, despite its promise, the authors note CASTER's limitations, including the absence of branch lengths and reliance on specific evolutionary model assumptions. Future developments aim to address these theoretical and practical challenges, expanding the tool's applicability to broader data types and more complex biological scenarios.
For reporters interested in research integrity issues, author Siavash Mir Arabbaygi notes, "Our field has made considerable strides in ensuring (almost) all tools used in practice are open source (including the method introduced in this paper and all other methods introduced by our lab). Top journals in the field also do a great job of nudging authors to provide data in public repositories such as Dryad, Zenodo, and FigShare. Authors provide data at various levels of detail, partially due to the effort needed to export large datasets and partially because of several limitations that public repositories have on the size of the provided data. Unlike some other fields, phylogenetics research has been very open and generous in terms of data sharing."
