Annotated genome assemblies for Geoscapheus dilatatus, Panesthia cribrata and Neogeoscapheus hanni

Genetic changes that enabled the evolution of eusociality have long captivated biologists. More recently, attention has focussed on the consequences of eusociality on genome evolution. Studies have reported higher molecular evolutionary rates in eusocial hymenopteran insects compared with their solitary relatives. To investigate the genomic consequences of eusociality in termites, we analysed nine genomes, including newly sequenced genomes from three non-eusocial cockroaches. Using a phylogenomic approach, we found that termite genomes have experienced lower rates of synonymous substitutions than those of cockroaches, possibly as a result of longer generation times. We identified higher rates of non-synonymous substitutions in termite genomes than in cockroach genomes, and identified pervasive relaxed selection in the former (24–31% of the genes analysed) compared with the latter (2–4%). We infer that this is due to reductions in effective population size, rather than gene-specific effects (e.g. indirect selection of caste-biased genes). We found no obvious signature of increased genetic load in termites, and postulate efficient purging of deleterious alleles at the colony level. Additionally, we identified genomic adaptations that may underpin caste differentiation, such as genes involved in post-translational modifications. Our results provide insights into the evolution of termites and the genomic consequences of eusociality more broadly.

Methods
Three Blaberidae genomes (Geoscapheus dilatatus, Panesthia cribrata and Neogeoscapheus hanni) were sequenced and assembled to investigate the evolution of this group, and to provide genomic resources for studies on Blattodea. These genomes were assembled using a combination of linked-read, long-read and Hi-C data (the raw seqeunce data are avalable on the SRA database under BioSample accessions SAMN39450771, SAMN39450772 and SAMN39450773 for Geoscapheus dilatatus, Neogeoscapheus hanni and Panesthia cribrata respectively). Assembly methods are outlined in the associated manuscript. Briefly, an initial de novo Geoscapheus dilatatus genome assembly was generated using the program Supernova with linked-read stLFR data. Gaps were filled using the program TGS-GapCloser with low-coverage long-read PacBio HiFi data, and the assembly was subsequently scaffolded using the SALSA2 with Hi-C reads.

An initial Panesthia cribrata genome assembly was generated using the program Supernova with linked-read TELL-Seq data. A second de novo assmembly was  generated using the program IPA with long-read PacBio Hifi data. The two assemblies were then merged using the program QuickMerge.

The Neogeoscapheus hanni genome assembly was generated using the program Supernova with linked-read stLFR data.

All three genomes were annotated using FGENESH++.