Anchored phylogenomics and a revised classification of the planidial larva clade of jewel wasps (Hymenoptera: Chalcidoidea)

Abstract

Planidia are free-living, mobile first-instar larvae that are notable in their ability to transition across different larval stages of their host, and for completing their development on the host prepupa as ectoparasitoids, effectively acting as larvalpupal external koinobionts. Chalcidoid taxa with a planidium form a monophyletic group, the planidial larva clade (PLC). We conducted a phylogenomic study of the PLC using anchored hybrid enrichment data. Phylogenetic analyses support the backbone relationship of PLC as: (Eutrichosomatinae, ((Philomidinae, Chrysolampinae), (Perilampinae, Eucharitidae))). Although excluded from the main analyses, the genus Jambiya, based on only 11 loci recovered, was placed as the sister of Chrysolampinae + Philomidinae or Perilampinae + Eucharitidae. Our results demonstrate that Perilampidae (Philomidinae, Chrysolampinae and Perilampinae) are paraphyletic. Divergence dating based on four node calibrations based on fossils suggests that the PLC arose approximately 111 Ma and the evolution of ant parasitism at least 64 Ma. Host associations, direct versus indirect hyperparasitism, ability to attack a host within a cocoon, soft versus hard planidial forms and mobility of the planidium were explored using a likelihood-based ancestral state reconstruction method. A revised higherlevel classification of the PLC is proposed, with Eutrichosomatinae elevated to Eutrichosomatidae (stat. rev.), Chrysolampinae and Philomidinae placed in Chrysolampidae (stat. rev.), Perilampidae (stat. rev.) restricted to what was referred to as Perilampinae, and Eucharitidae maintained with four subfamilies, with Akapalinae (unknown biology) as sister group to the core Eucharitidae, all of which are ant parasitoids. Jambiya is treated as an incertae sedis taxon within the planidial clade.

The Chalcidoidea bush of life – a massive radiation blurred by mutational saturation

“…Unsurprisingly, given their extreme morphological disparity and rampant convergence, a major collaborative effort to provide a resolved morphological tree for the superfamily largely failed (233 morphological characters scored on 283 species in 19 families; , although several family-level groups were recovered that were not found in the analyses of ribosomal markers. Within Chalcidoidea, when phylogenomic studies focused on smaller taxonomic units, the results were either in strong agreement with morphology, behavior or biogeographic hypotheses (e.g., Rasplus et al 2020), or strongly conflicting on some areas of the tree with intuitive and previously supported hypotheses (e.g., Cruaud et al 2021;Zhang et al 2022). When there is conflict, there is always the possibility that properties of the genomic data or confounding signal in morphological/biological data may be affecting the results.…”

“…Exons -Exons were obtained for 520 taxa following the protocol in Zhang et al (2022), of which 51 taxa were retrieved from previously published transcriptomes or genomes (Peters et al 2018;Zhang et al 2020), and 469 taxa (363 used in combined dataset) were enriched using the anchored hybrid enrichment (AHE) probe sets (Hym_Ich set or Hym_Cha set). See supplemental methods for details on library preparation and sequencing.…”

“…Exons -Assembly of loci, orthology assessment and contamination checking for the AHE520 data set followed Zhang et al (2022) (see also supplementary methods). Exons were selected of 414 taxa that were compatible for data combination with the UCE taxa, either as the same extraction, same species, same genus, or in 3 cases as closely related genera (Table S1b).…”

“…Exons were selected of 414 taxa that were compatible for data combination with the UCE taxa, either as the same extraction, same species, same genus, or in 3 cases as closely related genera (Table S1b). However, the assembly protocol used in Zhang et al (2022) was overly stringent, resulting in 28.5% of missing or ambiguous nucleotide calls. Therefore, raw sequence data for these 414 taxa were re-processed with HybPiper to increase completeness (Johnson et al 2016).…”

“…While genomic data offer great promise to resolve the tree of life, analytical challenges must be overcome. When more markers are analyzed, the probability of observing conflicting signal among them increases (Kumar et al 2012; Philippe et al 2017; Zhang et al 2020; Zhang et al 2022). Highly supported trees can be inferred that, without feedback from taxonomists, are considered to be accurate even though inferences may be flawed (Wiens 2004; Zhang et al 2022).…”