Toward a paradigm shift in comparative phylogeography driven by trait-based hypotheses

Abstract

For three decades, comparative phylogeography has conceptually and methodologically relied on the concordance criterion for providing insights into the historical/biogeographic processes driving population genetic structure and divergence. Here we discuss how this emphasis, and the corresponding lack of methods for extracting information about biotic/intrinsic contributions to patterns of genetic variation, may bias our general understanding of the factors driving genetic structure. Specifically, this emphasis has promoted a tendency to attribute discordant phylogeographic patterns to the idiosyncracies of history, as well as an adherence to generic null expectations of concordance with reduced predictive power. We advocate that it is time for a paradigm shift in comparative phylogeography, especially given the limited utility of the concordance criterion as genomic data provide ever-increasing levels of resolution. Instead of adhering to the concordance-discordance dichotomy, comparative phylogeography needs to emphasize the contribution of taxon-specific traits that will determine whether concordance is a meaningful criterion for evaluating hypotheses or may predict discordant phylogeographic structure. Through reference to some case studies we illustrate how refined hypotheses based on taxon-specific traits can provide improved predictive frameworks to forecast species responses to climatic change or biogeographic barriers while gaining unique insights about the taxa themselves and their interactions with their environment. We outline a potential avenue toward a synthetic comparative phylogeographic paradigm that includes addressing some important conceptual and methodological challenges related to study design and application of model-based approaches for evaluating support of trait-based hypotheses under the proposed paradigm.

Keywords: Pleistocene; biogeography; coalescent modeling; ecological traits; statistical phylogeography.

Fig. 1.

Generic vs. refined hypotheses to test the role of sea-level changes in driving diversification of island taxa. (A) A generic hypothesis assumes simultaneous divergence across taxa with disparate ecological traits and dispersal abilities. (B) A refined hypothesis limits the expectations of concordance to taxa subject to similar levels of population connectivity and persistence based on their habitat preference (top beetles in black) and excludes closely related taxa subject to high local extinction rates that would supersede divergence driven by past sea-level change (lower beetles in gray). Temporal concordance in population divergence among ecologically similar taxa provides support for the role of island connectivity cycles in driving diversification. At the same time, discordance between sets of similar taxa with different habitat preferences identifies habitat stability as a key factor structuring genetic variation in this island system (23).

Fig. 2.

The iDDC procedure generates species-specific predictions for patterns of genetic variation, which when coupled with approximate Bayesian computation (ABC), can be used to test biologically informed hypotheses about the effect of taxon-specific traits on phylogeographic structure. For example, it has been used to address questions on how the effects of climate change differ between wet- vs. dry-adapted montane sedges from the Rocky Mountains (24). The two closely related and codistributed Carex species have very similar distributional models (A), based on environmental niche modeling (26). However, when considering microhabitat affinity, different carrying capacities are predicted across the glaciated area at LGM (B), and coalescent simulations produce distinct expectations for patterns of genetic variation across the spatiotemporally dynamic landscape (C). Model comparisons using ABC supported a barrier model (i.e., zero carrying capacity in the glaciated region at LGM) for the wet-adapted species, which was presumably displaced during glacial periods due to accumulation of snow in wet microhabitats (e.g., drainages). On the contrary, a permeable model (i.e., allowing for nonzero carrying capacity in the glaciated region) is supported for the dry-adapted species, which could have persisted in situ within the glaciated areas, as drier microhabitats (e.g., ridges) remained relative free of persistent snow (24). Photos reproduced from ref. with permission of the publisher.