Population genetics of sponges in marine lakes
The use of thousands of RADseq-based SNPs provided the resolution necessary to reveal genetic patterns of Suberites diversicolorthat had not previously been captured at finer spatial scales. We observed clear clustering for the marine lake locations per lake. The lagoon populations Bay and to a lesser extend DAR showed to hold a basal position in the phylogenetic tree and in PCAs. They also showed links to most other populations in the migration network. The presence of ancestral polymorphisms in the marine lake populations could explain this pattern. The observation of finding more structure when using higher numbers of genetic markers has been shown in other marine organisms as well (Bradbury et al. , 2015; Maas et al. , 2018; Lemopoulos et al. , 2019; D’Aloia et al. , 2020; Sundeet al. , 2020; Timm, 2020). In a comparison among three high-throughput genotyping approaches, the RADseq generated markers were found to be the most sensitive and robust in detecting fine-scaled structure (D’Aloia et al. , 2020). The discrepancy in observed genetic structure based on a higher number of markers as compared to single markers is important in interpreting results from other studies for sessile marine organisms using low resolution markers.
While traditional, low-resolution markers have been useful in exposing morphologically cryptic sponge species, they have often failed to detect within-species diversity (as reviewed in Oppen et al., 2002; Pérez-Portela & Riesgo, 2018; Uriz & Turon, 2012; Wörheide et al., 2005). Using the high resolution of RAdseq generated markers allowed us to see clear clustering per lake even on very small spatial scales 1-10km. The scale at which we find strong structure is smaller compared to studies using microsatellites in the sponges Crambe crambe(Duran et al. , 2004), Scopalina lophyropoda (Blanquer and Uriz, 2010), Spongia lamella (Noyer and Becerro, 2012; Pérez-Portela, Noyer and Becerro, 2015), Stylissa carteri (Gileset al. , 2015), Cliona delitrix (Chaves-Fonnegra et al. , 2015), Xestospongia muta (Richards et al. , 2016),Paraleucilla magna (Guardiola, Frotscher and Uriz, 2016), Plenaster cragi (Taboada et al. , 2018) and Petrosia ficiformis (Riesgo et al. , 2019). Even studies using higher resolution markers also little structure at small spatial scales, with Brown et al. (2017) detecting little structuring forAphrocallistes vastus in British Colombia at scales <275km and Leiva et al. (2019) finding panmixia at scales >900km for Dendrilla antarctica . It could be that these are highly connected populations, possibly through rafting or sperm-mediated gene flow (Maldonado, 2006; DeBiasse, Nelson and Hellberg, 2014). Yet it is also possible that the number of SNPs from Brown et al. (2017) and Leiva et al. (2019) (67 and 529, respectively) was too low to detect subtle structure at small scales. Alternatively, the filtering strategy of these studies possibly was not rigorous enough to eliminate sufficient or all microbial contamination, possibly clouding patterns.
We assessed the effects of several drivers of population diversity and structure. First, we tested to what extent marine genetic differentiation conforms to the decay of population similarity with geographical distance resulting in a pattern of isolation‐by‐distance (Wright, 1943) using only Lineage B. We found strong population structure with clustering per lake, yet no pattern of isolation‐by‐distance was observed. This is remarkable, since we sampled at geographical distances of 1km - 1,400km. We also did not detect a pattern of isolation-by-environment, despite the great environmental variability among lakes (temperature: 29 - 32.4 °C, salinity: 24 - 33.4 ppt). Previous studies using a low number of markers did find a pattern of isolation-by-distance for sponges (Duran et al. , 2004; Blanquer and Uriz, 2010; Noyer and Becerro, 2012; Pérez-Portela, Noyer and Becerro, 2015), which is usually expected for species with restricted dispersal abilities (Wörheide, Solé-Cava and Hooper, 2005; Maldonado, 2006). Other studies report an influence of oceanographic currents (Chaves-Fonnegra et al. , 2015; Richards et al. , 2016; Riesgo et al. , 2019), or environmental heterogeneity (temperature and productivity) (Giles et al. , 2015) on sponges. Our results indicate that mechanisms other than only dispersal limitation by geographical distance or local environments are important in structuring S. diversicolor populations. In addition, the permeability of the landscape barrier surrounding the marine lakes, determining the degree of water flowing in and out of the lakes, did not seem to influence the population structure or diversity. PerhapsS. diversicolor populations are truly isolated per lake as their low dispersal ability restricts effective gene flow. Then, populations can become differentiated through genetic drift or via local adaptation to environmental parameters that we have not recorded (Frankham, Briscoe and Ballou, 2002). Alternatively, founder effects and subsequent priority effects could explain the pattern (Orsini et al. , 2013; Fukami, 2015; De Meester et al. , 2016).
Priority effects were previously discussed as potential drivers of structure in marine lake organisms by Maas et al. (2018) and de Leeuw et al. (2020). Depending on spatial scale Maas et al. (2018) found an effect of geographic distance and connectivity influencing mussel population structure. They argued that despite founder events stochastically driving alleles to fixation in small populations, ongoing dispersal would overwhelm this effect (Mayr, 1963; Waters, Fraser and Hewitt, 2013). Mussels have extensive pelagic larval duration periods, and Maas et al. (2018) hence argued that priority effects mediated by local adaptation could facilitate the observed patterns of population structure (Orsini et al. , 2013; Fukami, 2015; De Meester et al. , 2016). Sponges, in contrast, generally have poor dispersal abilities (Maldonado, 2006). As the current study does not find an effect of connection to the sea in structuring populations, stochastic fixation of alleles due to genetic drift may be the cause of each population being distinct. Including more lakes with replicates of local environments and/or connection to the sea may further elucidate drivers of sponge differentiation in fragmented habitats.