Dimensions of genotypic and phenotypic divergence in marine metapopulations: space, time, and ontogeny

Abstract

The tussle between selection and gene flow has considerable implications for evolution. As directly opposing counterparts, divergent selection should drive differentiation among subpopulations, whilst gene flow homogenises subpopulations. The balance between these two forces determines the evolutionary trajectories within and among subpopulations. However, many different outcomes can manifest, contingent on eco-evolutionary conditions, which makes predicting the product of selection–gene flow interactions challenging. Two particularly intriguing considerations are adaptive evolution amidst high gene flow and variability in selection–gene flow interactions.Under strong gene flow, considerable selection is required to overcome homogenisation. Successful adaptation is predicted to produce genotype-phenotype-environment associations. However, phenotypic plasticity might be favoured under high gene flow because it facilitates phenotype-environment matching when there is appreciable likelihood for genotype-environment mismatches. Species with extensive dispersal therefore represent interesting systems to investigate the conditions that local adaptation emerges, and contributions of directly heritable and plastic components to phenotypes, amidst homogenising gene flow.Additionally, factors affecting the pattern of selection and gene flow are not static, and selection–gene flow interactions are potentially variable across different scales. Spatial and temporal variation might exist in both these forces. Moreover, dispersal might be limited to specific life stages, or selection may not be consistent across an individual’s lifetime. Collectively, the ecological dimensions of space, time, and ontogeny have potentially complex ways of structuring genotypic and phenotypic variation within and among subpopulations. Understanding when, where, and how this variation is structured can elucidate the effect of selection–gene flow interactions on biodiversity and genotype-phenotype-environment associations in natural settings.Marine species are extremely useful for studying how space, time, and ontogeny structure biological variation in high gene flow systems. Many marine species occupy discrete habitat patches in metapopulation-like systems. Considerable dispersal potential affords high connectivity, yet dynamic marine environments cause significant spatial and temporal heterogeneity in processes that modulate gene flow and selection. A vast majority of marine organisms also have complex life cycles, which sets the stage for possible selective antagonism across ontogeny, and dispersal that is restricted to early life stages (particularly in benthic taxa).In this thesis, I utilise Coco’s frillgoby, Bathygobius cocosensis, to study how space, time, and ontogeny structure genotypic and phenotypic divergence in marine metapopulations. Through explicit interrogation across multiple ecological dimensions and a combination of diverse datasets (life history traits, morphology, and genome-wide SNPs), this work paints a novel and comprehensive picture of how heterogeneity in selection and gene flow might affect the evolutionary outcomes in thalassic taxa.In Chapter 2, I investigate temporal variability in early life history covariances. This work utilised otolith-derived life history traits in juvenile B. cocosensis to test the hypothesis that complex life cycles partition trait variation between developmental phases. I found mixed support for this evolutionary expectation, suggesting that trade-offs might exist for different traits. I also tested for temporal differences in covariance patterns among cohorts as inferential evidence of variability in genotypic or environmental variation across time. My surprising observation of non-significant temporal differences in trait covariances suggests that processes contributing toward population-level life history phenotype distributions might be consistent through time.In Chapter 3, I collected spatial and temporal replicates of adult and juvenile B. cocosensis subpopulations to investigate life stage differences in genotypic structure. I contrasted genome-wide and locus-specific divergence to assess potential contributions of neutral and selective processes (respectively) to genotypic structure. My results were consistent with extensive gene flow, but also evidenced the presence of chaotic genetic patchiness, suggesting ephemerality between structured and well-connected states in the metapopulation. Furthermore, I also observed genetic signatures of temporally consistent selection that differed between life stages. Additionally, I studied head morphology (a putatively ecologically important trait) to quantify phenotypic structure among adult subpopulations. Despite considerable temporal turnover in local head shape, there was evidence of temporally consistent phenotypic structure among subpopulations, amidst evidence of considerable gene flow. These results suggest that high gene flow and heterogeneous environments may not limit adaptive potential in B. cocosensis and provides additional support for the role of selective processes in structuring subpopulations.In Chapter 4, I delved deeper into the scale of adaptation in B. cocosensis. I utilised a large number of individually genotyped (SNPs) and phenotyped (head shape morphology) fish to determine the spatial and temporal scale that genotype-phenotype-environmental associations manifest in high gene flow systems. Among subpopulations (separated through time and space), I found evidence of significant phenotypic structuring, despite panmixia, and detected two SNPs with putative genotype-phenotype associations. Within a subpopulation, across tide pools with different microhabitats, I also observed significant phenotype-environment associations. These results support the role of environmental heterogeneity in structuring phenotypic variation at fine and broad spatial scales in B. cocosensis. Plastic responses likely play a considerable role in generating phenotypic variation but these results also suggest a partial genetic basis to head shape morphology.