TY  - JOUR
T1  - Gene-Tree Conflict, Missing Data, and Model Misspecification Demand Explicit Uncertainty Rather Than False Certainty in Insect Phylogenomics
A1  - Chen Hao
A1  - Liu Fang
A1  - Zhao Lin
A1  - Wei Zhang
JF  - Entomological Research Letters
JO  - Entomol Res Lett
SN  - 3108-639X
Y1  - 2024
VL  - 4
IS  - 1
DO  - 10.51847/4z2DKQ8g8a
SP  - 146
EP  - 155
N2  - Phylogenomic datasets can produce apparently decisive insect phylogenies while retaining substantial uncertainty about the histories represented by individual loci, the adequacy of sampled data, and the assumptions linking gene trees to species trees. This article develops a framework for distinguishing uncertainty generated by biological gene-tree heterogeneity, missing taxa or loci, orthology error, model misspecification, and downstream species-tree inference. The central argument is that conventional branch support cannot function as a universal summary of these distinct uncertainties. Gene-tree discordance may represent incomplete lineage sorting, introgression, duplication and loss, or analytical error; missingness may interact with taxon representation and phylogenetic information rather than operate as a simple percentage threshold; and apparent orthology can be disrupted by differential gene loss without making all multicopy data unusable. Model adequacy introduces a further boundary between strongly supported inference and adequately represented evolutionary processes. The proposed uncertainty-source architecture therefore separates the origin of uncertainty, its observable manifestation, the inference stage it can affect, and the degree to which it remains unresolved after species-tree estimation. The framework does not assign numerical certainty grades or identify a single correct topology. Instead, it establishes a basis for reporting when confidence is localized to a statistical procedure while biological or analytical alternatives remain unresolved. Explicitly preserving those alternatives should make insect phylogenomic conclusions more interpretable, reproducible, and appropriately conditional.
UR  - https://esvpub.com/article/gene-tree-conflict-missing-data-and-model-misspecification-demand-explicit-uncertainty-rather-than-qyhzgilsifk0w6t
ER  -