Regulatory RNA Patterns Associated with Bee Evolution Should Be Interpreted with Caution, but Correlational Studies Should Not Be Entirely Dismissed

Abstract

In this response to comments by Heneberg (2026) on our recent review, “The evolution of bees: insights from ‘omic studies,” we continue the scientific conversation about evidence linking RNA regulatory mechanisms to phenotypic traits. We acknowledge that studies connecting small noncoding RNAs and alternative splicing to social behavior phenotypes are largely correlational and should be interpreted with an appropriate level of caution. At the same time, we suggest that such findings may still offer valuable insights, given the hypothesized importance of regulatory mechanisms in social evolution and the substantial body of work in model organisms demonstrating causal links between these processes and phenotypic change.

Significance.

In this letter, we respond to Heneberg (2026) to continue the scientific discourse on the cautious interpretation of studies of alternative splicing, long noncoding RNAs, microRNAs, and circular RNAs and their potential effects on traits. Many of the studies tying social phenotypes to small RNAs and alternative splicing have thus far been largely correlational and require caution with their interpretation. Nonetheless, there is value in discussing the possible connection between these mechanisms and social phenotypes, as previous studies have demonstrated that such mechanisms can causally influence traits in other organisms.

As part of our recent review on bee genomics and biology (Brenman-Suttner and Zayed 2026), we summarized some of the recent research highlighting associations between regulatory mechanisms such as alternative splicing, long noncoding RNA, circular RNA, and microRNA with social phenotypes in bees. We were particularly interested in reviewing this literature, given the multitude of hypotheses (Amdam et al. 2004; Toth and Robinson 2007; Page et al. 2012; Kapheim et al. 2015, 2020; Toth and Rehan 2017; Rubenstein et al. 2019) that emphasize the role of gene regulation and genomic regulatory architecture in eusocial evolution. In their commentary, Heneberg (2026) provides a critical analysis of the studies we reviewed and insightfully notes that most of this research is correlational and should be interpreted with caution.

We agree with Heneberg (2026). We are fully aware that typical expression profiling studies in bees are not appropriate for inferring causality, as one of us has discussed in previous reviews (Zayed and Robinson 2012; Toth and Zayed 2021). So, while our summary of the recent literature in this section was written with some rhetorical enthusiasm, we often used cautious and noncausal language when discussing this body of work. For example, we used modifiers, such as “may suggest,” “associated”, “linked”, “highlighted their connection”, or “may have contributed”. We did not explicitly use the word “cause” or “causal” in this section. Alas, one limitation of studying bees is that we typically lack the sophisticated toolkits for genetic manipulation found in genetically tractable model organisms.

Nevertheless, we were keen to highlight these intriguing “associations” because of both the hypotheses that stipulate that regulatory mechanisms “underlie” key eusocial phenotypes and due to the manipulative experiments that have causally linked these regulatory mechanisms to important phenotypes in some insects. For example, alternative splicing is causally involved in complementary sex determination in honey bees (Hasselmann et al. 2008; Seiler and Beye 2024). RNAi-based manipulations have linked specific splice variants to traits such as ovary activation and social parasitism in honeybees (Jarosch et al. 2011), insecticide resistance in the carmine spider mite Tetranychus cinnabarinus and the fruit fly Drosophila melanogaster (Wei et al. 2023), sex determination in the sawfly Athalia rosae (Mine et al. 2021) as well as the red flour beetle Tribolium castaneum (Shukla and Palli 2012), and cold acclimation in D. melanogaster (Hopkins et al. 2018). Similarly, experimental manipulation of miRNA expression via overexpression or inhibition has shown that miRNAs can causally influence many traits across a diverse group of species. These experiments demonstrate that knocking down or overexpressing a particular small RNA leads to an altered phenotype for the trait in question, showing its causality (Reynolds 2024). This has been shown in relation to molting in the sugarcane shoot borer Chilo infuscatellus (Wang et al. 2025), wing development in both the brown planthopper Nilaparvata lugens (Tian et al. 2025) and the brown citrus aphid Aphis citricidus (Shang et al. 2020), the timing of development as seen in both the diamondback moth Plutella xylostella (Bardapurkar et al. 2025) as well as the oriental fruit fly Bactrocera dorsalis (Zhang et al. 2020b), and insecticide resistance in P. xylostella (Etebari et al. 2018). Similarly, RNAi- and siRNA-based knockdown experiments have demonstrated that circRNAs can directly influence traits such as antiviral defense in the small brown planthopper Laodelphax striatellus (Zhang et al. 2020a), development of ovaries in the Asian tiger mosquito Aedes albopictus (Gao et al. 2023), and wing morphology in the brown citrus aphid Aphis citricidus (Liu et al. 2025). While some circRNAs act through mechanisms that do not rely on acting as a sponge for miRNAs (Zhang et al. 2020a), other studies provide causal evidence that circRNAs function as competing endogenous RNAs, in which circRNA knockdown, miRNA overexpression, or knockdown of downstream target genes result in the same phenotypic outcomes (Gao et al. 2023; Liu et al. 2025). Finally, RNAi-based knockdown experiments have shown that lncRNAs can causally influence insecticide resistance in the tobacco cutworm Spodoptera litura (Shi et al. 2022) and hormone-regulated developmental processes, including autophagy in the silkworm Bombyx mori (Qiao et al. 2021) and metamorphosis of pupae in the flesh fly Sarcophaga peregrina (Shang et al. 2023).

We look forward to future studies that go beyond correlation as our community embraces new tools and techniques for genetic manipulation in nongenetically tractable organisms to give a clearer understanding of the mechanistic relationship between alternative splicing, non-mRNA molecules, and bee social behavior.

Funding

The research was funded by a Discovery Grant from The Natural Sciences and Engineering Research Council of Canada (NSERC), and a York University Research Chair in Genomics to A.Z.

Data Availability

No new data were generated or analyzed in support of this research.