Systematic bias undermines mitochondrial phylogenomics of deep divergences of Diptera
- 1Bioengineering Research Laboratory, Department of Biosciences and Bioengineering, Indian Institute of Technology Guwahati
- 2Centre for the Environment, Indian Institute of Technology Guwahati
Abstract
Mitochondrial genomes (mtDNA) are widely used in insect phylogenetics, yet their reliability for resolving ancient and rapid radiations, such as in Diptera, remains confounded by systematic bias. We analyzed 112 dipteran mitogenomes, treating the 13 protein-coding genes as partitions of a single non-recombining locus, to quantify the limits of mitochondrial signal and identify sources of deep-branch inference error. While mtDNA retains a structured, largely tree-like signal that reliably resolves terminal clades, deeper backbone relationships remain unstable and inconsistent across inference methods. This instability arises from compositional non-stationarity, substitution saturation, and lineage-specific rate heterogeneity; when these properties violate the assumptions of homogeneous models, they generate misleading statistical support for incorrect topologies. Although heterogeneous methods, including reduced-state recoding and heterotachy-aware mixture models, alleviate some biases, deep-node distortions persist. Diptera-parameterized simulations successfully reproduce these empirical patterns through compositional convergence, covarion-like heterotachy, and terminal rate acceleration, directly linking specific evolutionary processes to inference error. Extensive conflict among the 13 non-recombining partitions, negative internode certainty, and spurious non-tree-like signals further diagnose severe systematic bias. Tree-based metrics reveal predictable structural distortions, including branch-length inflation and reduced stemminess. Together, our results show that dipteran mitogenomes are information-rich yet analytically fragile at depth, and we provide a diagnostic workflow demonstrating that apparent phylogenetic complexity can emerge from structured error and homoplasy rather than true evolutionary history.
Related
- Type-Strain-Anchored Phylogenomics Reveals Taxonomic Discordance and Functional Variation in Agrobacterium
- Phylogenomics Illuminates the Evolutionary History of Wild Silkmoths in Space and Time (Lepidoptera: Saturniidae)
- The genome sequence of a robber fly, Cyrtopogon pulchripes Loew, 1871 (Diptera: Asilidae)
- The genome sequence of the Coltsfoot Fly, Acidia cognata (Wiedemann, 1817) (Diptera: Tephritidae).
- Molecular and functional characterization of a typical 2-Cys peroxiredoxin and its developmental and pyriproxyfen-responsive expression in Aedes albopictus (Diptera: Culicidae)
BibTeX
@article{kabiraj2026systematic,
title = {{Systematic bias undermines mitochondrial phylogenomics of deep divergences of Diptera}},
author = {Debajyoti Kabiraj and Pulakeswar Basumatary},
journal = {Scientific Reports},
year = {2026},
month = aug,
doi = {10.1038/s41598-026-63009-w},
url = {https://doi.org/10.1038/s41598-026-63009-w},
note = {CC BY},
}
RIS
TY - JOUR
TI - Systematic bias undermines mitochondrial phylogenomics of deep divergences of Diptera
AU - Debajyoti Kabiraj
AU - Pulakeswar Basumatary
PY - 2026
DA - 2026/08/01
JO - Scientific Reports
DO - 10.1038/s41598-026-63009-w
UR - https://doi.org/10.1038/s41598-026-63009-w
AB - Mitochondrial genomes (mtDNA) are widely used in insect phylogenetics, yet their reliability for resolving ancient and rapid radiations, such as in Diptera, remains confounded by systematic bias. We analyzed 112 dipteran mitogenomes, treating the 13 protein-coding genes as partitions of a single non-recombining locus, to quantify the limits of mitochondrial signal and identify sources of deep-branch inference error. While mtDNA retains a structured, largely tree-like signal that reliably resolves terminal clades, deeper backbone relationships remain unstable and inconsistent across inference methods. This instability arises from compositional non-stationarity, substitution saturation, and lineage-specific rate heterogeneity; when these properties violate the assumptions of homogeneous models, they generate misleading statistical support for incorrect topologies. Although heterogeneous methods, including reduced-state recoding and heterotachy-aware mixture models, alleviate some biases, deep-node distortions persist. Diptera-parameterized simulations successfully reproduce these empirical patterns through compositional convergence, covarion-like heterotachy, and terminal rate acceleration, directly linking specific evolutionary processes to inference error. Extensive conflict among the 13 non-recombining partitions, negative internode certainty, and spurious non-tree-like signals further diagnose severe systematic bias. Tree-based metrics reveal predictable structural distortions, including branch-length inflation and reduced stemminess. Together, our results show that dipteran mitogenomes are information-rich yet analytically fragile at depth, and we provide a diagnostic workflow demonstrating that apparent phylogenetic complexity can emerge from structured error and homoplasy rather than true evolutionary history.
SN - 2045-2322
ER -