Genomic insights from natural history collections reveal cryptic speciation in coral-guard crabs (family: Trapeziidae)

Mutualistic relationships such as the one between Trapezia crabs and coral colonies are common in reef organisms and play a crucial role in coral resilience and resistance to climate-induced stressors, yet very little is known about the taxonomic diversity and evolutionary history of the species involved. Despite being essential actors of coral reefs and threatened by the ongoing degradation of their habitat, little genetic information is available for Trapezia crabs, including the exact number of species and their relationships. To overcome this limitation, we sampled Natural History collections, an important and underutilized source of genomic data. We used a novel approach optimized for degraded DNA, to generate high-quality genomic data from a combination of 166 museum tissues and freshly collected samples and recovered a strongly supported phylogeny of the Trapezia genus, clarifying species relationships of a majority of taxa and suggesting the potential division of Trapezia into two genera. We then focused on the most widespread species T. bidentata and identified four distinct genetic clusters suggesting high divergence and cryptic speciation in the Indian Ocean and the Marquesas Islands. Populations of the Central and West Pacific, showed signs of admixture across a heterogeneous seascape, attributing to a potentially long pelagic dispersal phase and expansive gene pool. Our results highlight the need to further explore the genetic diversity within other Trapezia species and other coral-associated organisms, as they are likely to exhibit more complex genetic patterns than previously understood.

Methods

Samples were collected from the reefs of Guam and museum repositories from the Florida Museum of Natural History and the Paris Museum National d’Histoire Naturelle. DNA was extracted and sent for Genotyping by Random Amplicon Sequencing-Direct (GRAS-Di®, Enoki & Takeuchi 2019) sequencing at the University of California, Davis. Reads were processed and SNP matrices were created in STACKS v2.60 (Catchen et al. 2011, 2013).