10.1093/sysbio/syu131

Summary

Ingested 2026-04-22. 2 findings extracted and verified.

Findings worth citing

Finding 1 — On the empirical cetacean phylogeny, more than 77% of neutral character datasets simulated without any state-dependent diversification showed a statistically significant association between character state and speciation rate (p < 0.05) under BiSSE.

More than 77% of the 400 character sets showed a significant (p < 0.05) association between character state and speciation rate, and 58% rejected the character-independent model with great confidence (p < 0.001). — p. 10

Why this is citable: This finding directly quantifies the Type I error rate of BiSSE on a real phylogeny, providing a concrete benchmark for how unreliable trait-dependent diversification inferences can be — essential to cite when evaluating any BiSSE-based result.

Counter / limitation: This result is specific to the cetacean phylogeny, which has well-documented heterogeneous diversification (dolphin radiation); the magnitude of false positives may be lower on phylogenies with more homogeneous diversification dynamics.

Topics: BiSSE, type_I_error, trait_dependent_diversification, phylogenetic_comparative_methods, model_inadequacy

Finding 2 — Taxon name length — a purely arbitrary character with no biological meaning — showed a significant association with speciation rate in more than 69% of vertebrate subtrees, approaching 100% for ray-finned fishes.

For all groups, more than 69% of trees showed a significant (p < 0.05) correlation between taxon name length and speciation rate; for fishes, this approached 100% (60 of 61 subtrees). — p. 17

Why this is citable: This result is maximally persuasive as a demonstration of BiSSE’s vulnerability, because taxon name length cannot conceivably cause speciation, making the false positive rate impossible to attribute to any biological confound.

Counter / limitation: Taxon name length does carry weak phylogenetic signal (congeners share name prefixes), so the test is not a fully clean null; some of the elevated error rate may reflect this residual phylogenetic structure rather than purely the tree’s diversification heterogeneity.

Topics: BiSSE, type_I_error, trait_dependent_diversification, phylogenetic_comparative_methods, model_inadequacy