Phylogenomic data exploration with increased sampling provides new insights into the higher‐level relationships of butterflies and moths (Lepidoptera).

A robust and stable phylogenetic framework is a fundamental goal of evolutionary biology. As the third largest insect order, Lepidoptera (butterflies and moths) are central to terrestrial ecosystems and serve as important models for biologists studying ecology and evolutionary biology. However, for such an insect group, the higher‐level phylogenetic relationships among its superfamilies remain poorly resolved. Here, we increased taxon sampling among Lepidoptera (37 superfamilies and 68 families containing 263 taxa), obtaining a series of amino‐acid data sets from 69 680 to 400 330 aa in length for phylogenomic reconstructions. Using these data sets, we explored the effect of different taxon sampling with significant increases in gene loci on tree topology using maximum‐likelihood (ML) and Bayesian inference (BI) methods. Moreover, we also tested the effectiveness of topology robustness among the three ML‐based models. The results demonstrated that taxon sampling is an important determinant in tree robustness of accurate phylogenetic estimation for species‐rich groups. Site‐wise heterogeneity was identified as a significant source of bias, causing inconsistent phylogenetic positions among ditrysian lineages. The application of the posterior mean site frequency (PMSF) model provided reliable estimates for higher‐level phylogenetic relationships of Lepidoptera. Phylogenetic inference presented a comprehensive framework among lepidopteran superfamilies, and revealed some new sister relationships with strong supports (Papilionoidea is sister to Gelechioidea, Immoidea is sister to Galacticoidea, and Pyraloidea is sister to Hyblaeoidea, respectively). The current study provides essential insights for future phylogenomic investigations in species‐rich lineages of Lepidoptera and enhances our understanding on phylogenomics of highly diversified groups. [ABSTRACT FROM AUTHOR]