Bridging classical taxonomy and modern approaches: statistical learning and computational species delimitation in Tetrahymena (Alveolata, Ciliophora)
Year: 2026
Type: article
Abstract: Species delimitation in ciliates remains challenging due to their morphological simplicity and the frequent occurrence of cryptic diversity. In the genus Tetrahymena , the mitochondrial cytochrome c oxidase subunit I ( COI ) gene has emerged as a key DNA barcode marker, yet the genetic thresholds separating intra- from interspecific variation remain poorly defined. To address this issue, we analyzed all publicly available COI sequences for Tetrahymena and applied a supervised statistical framework for species delimitation. Using expert-derived taxonomic assignments, we inferred data-driven distance thresholds that optimally distinguish within-species from between-species comparisons and compared their performance with coalescent-based methods (PTP and mPTP). The supervised approach achieved high classification accuracy (>98%) and low misclassification rates (<2%), whereas mPTP tended to lump taxa. These findings indicate that data-driven threshold inference based on expert taxonomy provides a robust and scalable framework for species delimitation, improving the integration of molecular data into ciliate systematics and refining species boundaries in Tetrahymena .
Authors: Fabiola S. Costa(Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Software, Supervision, Validation, Writing – original draft, Writing – review & editing), Carlos G. Schrago(Data curation, Formal analysis, Investigation, Methodology, Software, Supervision, Validation, Writing – original draft, Writing – review & editing), Roberto Júnio P. Dias(Conceptualization, Data curation, Formal analysis, Investigation, Supervision, Validation, Writing – original draft, Writing – review & editing), Mariana F. Rossi(Conceptualization, Data curation, Formal analysis, Investigation, Supervision, Validation, Writing – original draft, Writing – review & editing)
Language: English
Domain: Life Sciences
Open Access status: hybrid
Awards: G4007487310, G9001388642
OpenAlex ID: W7233983355
