Constraints on geographic variation in fiddler crabs (Ocypodidae: Uca) from the western Atlantic

Classifying high-dimensional phenotypes with ensemble learning

Classification is a fundamental task in biology used to assign members to a class. While linear discriminant functions have long been effective, advances in phenotypic data collection are yielding increasingly high-dimensional datasets with more classes, unequal class covariances, and non-linear distributions. Numerous studies have deployed machine learning techniques to classify such distributions, but they are often restricted to a particular organism, a limited set of algorithms, and/or a specific classification task. In addition, the utility of ensemble learning or the strategic combination of models has not been fully explored.We performed a meta-analysis of 33 algorithms across 20 datasets containing over 20,000 high-dimensional shape phenotypes using an ensemble learning framework. Both binary (e.g., sex, environment) and multi-class (e.g., species, genotype, population) classification tasks were considered. The ensemble workflow contains functions for preprocessing, training individual learners and ensembles, and model evaluation. We evaluated algorithm performance within and among datasets. Furthermore, we quantified the extent to which various dataset and phenotypic properties impact performance.We found that discriminant analysis variants and neural networks were the most accurate base learners on average. However, their performance varied substantially between datasets. Ensemble models achieved the highest performance on average, both within and among datasets, increasing average accuracy by up to 3% over the top base learner. Higher class R2 values, mean class shape distances, and between- vs. within-class variances were positively associated with performance, whereas higher class covariance distances were negatively associated. Class balance and total sample size were not predictive.Learning-based classification is a complex task driven by many hyperparameters. We demonstrate that selecting and optimizing an algorithm based on the results of another study is a flawed strategy. Ensemble models instead offer a flexible approach that is data agnostic and exceptionally accurate. By assessing the impact of various dataset and phenotypic properties on classification performance, we also offer potential explanations for variation in performance. Researchers interested in maximizing performance stand to benefit from the simplicity and effectiveness of our approach made accessible via the R package pheble.

Aggregate Trait Evolvability and Macroevolution in Two Sister Species of the Bryozoan Stylopoma

The study of trait evolution in modular animals is more complicated than that in solitary animals, because a single genotype of a modular colony can express an enormous range of phenotypic variation. Furthermore, traits can occur either at the module level or at the colony level. However, it is unclear how the traits at the colony level evolve. We test whether colony-level aggregate traits, defined as the summary statistics of a phenotypic distribution, can evolve. To quantify this evolutionary potential, we use parent-offspring pairs in two sister species of the bryozoan Stylopoma, grown and bred in a common garden breeding experiment. We find that the medians of phenotypic distributions are evolvable between generations of colonies. We also find that the structure of this evolutionary potential differs between these two species. Ancestral species align more closely with the direction of species divergence than the descendent species. This result indicates that aggregate trait evolvability can itself evolve.

Population Divergence along a Genetic Line of Least Resistance in the Tree Species Eucalyptus globulus

The evolutionary response to selection depends on the distribution of genetic variation in traits under selection within populations, as defined by the additive genetic variance-covariance matrix (G). The structure and evolutionary stability of G will thus influence the course of phenotypic evolution. However, there are few studies assessing the stability of G and its relationship with population divergence within foundation tree species. We compared the G-matrices of Mainland and Island population groups of the forest tree Eucalyptus globulus, and determined the extent to which population divergence aligned with within-population genetic (co)variation. Four key wood property traits exhibiting signals of divergent selection were studied—wood density, extractive content, and lignin content and composition. The comparison of G-matrices of the mainland and island populations indicated that the G-eigenstructure was relatively well preserved at an intra-specific level. Population divergence tended to occur along a major direction of genetic variation in G. The observed conservatism of G, the moderate evolutionary timescale, and close relationship between genetic architecture and population trajectories suggest that genetic constraints may have influenced the evolution and diversification of the E. globulus populations for the traits studied. However, alternative scenarios, including selection aligning genetic architecture and population divergence, are discussed.

Evolution of morphological integration in the skull of Carnivora (Mammalia): Changes in Canidae lead to increased evolutionary potential of facial traits

Morphological integration refers to the fact that different phenotypic traits of organisms are not fully independent from each other, and tend to covary to different degrees. The covariation among traits is thought to reflect properties of the species' genetic architecture and thus can have an impact on evolutionary responses. Furthermore, if morphological integration changes along the history of a group, inferences of past selection regimes might be problematic. Here, we evaluated the stability and evolution of the morphological integration of skull traits in Carnivora by using evolutionary simulations and phylogenetic comparative methods. Our results show that carnivoran species are able to respond to natural selection in a very similar way. Our comparative analyses show that the phylogenetic signal for pattern of integration is lower than that observed for morphology (trait averages), and that integration was stable throughout the evolution of the group. That notwithstanding, Canidae differed from other families by having higher integration, evolvability, flexibility, and allometric coefficients on the facial region. These changes might have allowed canids to rapidly adapt to different food sources, helping to explain not only the phenotypic diversification of the family, but also why humans were able to generate such a great diversity of dog breeds through artificial selection.

The unusual case of the widely distributed fiddler crab Minuca rapax (Smith, 1870) from the western Atlantic: an exemplary polytypic species

A classic dilemma in taxonomy is distinguishing intraspecific from interspecific variation. In order to better comprehend the process of divergence and speciation, we examine morphological, genetic, developmental and behavioural variation among related fiddler crab populations from eastern North America, the Caribbean and South America. We chose geographically remote populations that appear related to Minuca rapax (Smith, 1870). First, using females from across the range of the species, we use geometric morphometric techniques to identify regional differences in carapace shape. Second, in the northern portion of the range, the Caribbean into the Gulf of Mexico, we report variation in the relationship between corporal size and cheliped length in males. Third, we examine the major components of the courtship waves produced by males from several locations in the western Gulf of Mexico. Fourth, we compare the structure of the gastric mill between different populations in the Gulf of Mexico, the Caribbean and the Atlantic Ocean. And, fifth, we use mitochondrial 16S rDNA and cytochrome oxidase subunit I as genetic markers to define the phylogeographic relationship among specimens from more than 20 populations. From these studies, we find discrete, distinct populations across the original range of the species. In particular, populations in the northern Gulf of Mexico appear to represent a lineage that has resulted from limited gene flow and sustained selection pressures. On the basis of the observed degree of divergence, it is apparent that some separated populations in M. rapax should be recognised as evolutionary significant units. The geographic range of these populations is consistent with the historical range for Minuca virens (Salmon & Atsaides, 1968), a putative species that otherwise cannot be consistently distinguished from M. rapax based on discrete external morphological characters. This study provides evidence for M. virens as an emergent but possibly not completely isolated subclade of the M. rapax species complex.

Approaches to Macroevolution: 2. Sorting of Variation, Some Overarching Issues, and General Conclusions

Approaches to macroevolution require integration of its two fundamental components, within a hierarchical framework. Following a companion paper on the origin of variation, I here discuss sorting within an evolutionary hierarchy. Species sorting-sometimes termed species selection in the broad sense, meaning differential origination and extinction owing to intrinsic biological properties-can be split into strict-sense species selection, in which rate differentials are governed by emergent, species-level traits such as geographic range size, and effect macroevolution, in which rates are governed by organism-level traits such as body size; both processes can create hitchhiking effects, indirectly causing the proliferation or decline of other traits. Several methods can operationalize the concept of emergence, so that rigorous separation of these processes is increasingly feasible. A macroevolutionary tradeoff, underlain by the intrinsic traits that influence evolutionary dynamics, causes speciation and extinction rates to covary in many clades, resulting in evolutionary volatility of some clades and more subdued behavior of others; the few clades that break the tradeoff can achieve especially prolific diversification. In addition to intrinsic biological traits at multiple levels, extrinsic events can drive the waxing and waning of clades, and the interaction of traits and events are difficult but important to disentangle. Evolutionary trends can arise in many ways, and at any hierarchical level; descriptive models can be fitted to clade trajectories in phenotypic or functional spaces, but they may not be diagnostic regarding processes, and close attention must be paid to both leading and trailing edges of apparent trends. Biotic interactions can have negative or positive effects on taxonomic diversity within a clade, but cannot be readily extrapolated from the nature of such interactions at the organismic level. The relationships among macroevolutionary currencies through time (taxonomic richness, morphologic disparity, functional variety) are crucial for understanding the nature of evolutionary diversification. A novel approach to diversity-disparity analysis shows that taxonomic diversifications can lag behind, occur in concert with, or precede, increases in disparity. Some overarching issues relating to both the origin and sorting of clades and phenotypes include the macroevolutionary role of mass extinctions, the potential differences between plant and animal macroevolution, whether macroevolutionary processes have changed through geologic time, and the growing human impact on present-day macroevolution. Many challenges remain, but progress is being made on two of the key ones: (a) the integration of variation-generating mechanisms and the multilevel sorting processes that act on that variation, and (b) the integration of paleontological and neontological approaches to historical biology.