Phylogenomics resolves the century-old ‘Zoraptera problem’: Zoraptera as the earliest diverging lineage of Polyneoptera

Wing evolution enabled insects to pioneer aerial niches, a key innovation driving their remarkable radiation. However, interordinal relationships among early-diverging winged insects remain contentious, largely owing to the enigmatic phylogenetic position of Zoraptera, a species-poor and morphologically cryptic order. Since its discovery over a century ago, the ‘Zoraptera problem’ has persisted as a crucial issue in insect evolution, as alternative placements imply fundamentally different scenarios for polyneopteran diversification and trait evolution. Here, we sequenced and assembled two zorapteran genomes, integrating them with public data to compile a dataset of 89 polyneopteran species and 23 outgroups. Using 1367 universal single-copy orthologous genes, we evaluated phylogenetic signal under different evolutionary models and data treatments. Model testing revealed that the prevailing Haplocercatafirst hypothesis (Zoraptera + Dermaptera as the earliest polyneopteran branch) collapses when among-site compositional heterogeneity is modelled to reduce systematic errors. Critically, under the optimal site-heterogeneous model, Zoraptera emerges as the earliest-diverging polyneopteran lineage, with Dermaptera + Plecoptera forming the sister clade Dermoplectopterida to all other Polyneoptera, a topology corroborated by multispecies coalescent and gene jackknifing analyses. Our findings highlight the necessity of multi-model approaches in insect phylogenomics, providing a robust framework for early winged insect diversification.