Gene flow and models of avian speciation in tropical mountains

Abstract

Eight decades after the modern synthesis, the role of gene flow in speciation remains contentious. Because gene flow is a homogenizing force and sister species frequently occur in allopatry, geographic isolation was assumed to be a requirement for the majority of speciation events for most of the 20th century. More recently novel theory, genome-wide DNA sequence data and computational tools have provided evidence to challenge this assumption, demonstrating speciation with gene flow is both plausible and more common than previously thought. These results have prompted a shift in emphasis from the geography of diverging populations to the interplay of underlying evolutionary mechanisms. Yet convincing empirical examples remain rare, and imprecise language blurs the distinction between classical models of sympatric or parapatric speciation and allopatric speciation followed by extensive secondary contact and hybridization. When and where—if ever—is speciation with gene flow likely to occur? What, exactly, does “speciation with gene flow” indicate about the constancy and timing of migration? Can our tools distinguish even distinguish among these parameters? Tropical mountains are the most species-rich terrestrial environment on earth, with significant geographic complexity and strong, temporally stable environmental gradients to generate selection. These features provide a powerful natural laboratory to address fundamental questions in speciation, but research effort has historically lagged behind studies of temperate regions due to obstacles of access and cost. In this dissertation I use empirical data, theory, and evolutionary simulations to explore the role of gene flow during speciation events in tropical mountains. In Chapter 1, I validate a method for collecting genome-wide DNA sequence data from historical museum specimens and develop a bioinformatic pipeline to assemble loci and call variants. I apply this approach to degraded DNA from historical museum samples of a poorly known New Guinea kingfisher species, Syma torotoro, inferring population genetic structure in contiguous forest habitats. This methodology allows me to overcome logistical difficulties inherent to work in remote tropical environments to achieve appropriate sample sizes. In Chapter 2, I combine these data with additional samples from its montane sister species, Syma megarhyncha, and test a hypothesis of speciation with gene flow across an elevational gradient in Syma. I find evidence of assortative mating the face of extensive historical and contemporary gene flow, suggesting selection across mountainsides can maintain species limits in the face of incomplete reproductive isolation. In Chapter 3, I use theory and evolutionary simulations to explore the probability and genomic signature of speciation with alternating periods of isolation and gene flow. I find speciation with even brief lulls in gene flow is significantly easier than speciation with gene flow every generation and difficult to distinguish by standard methods. However, I also find periodic gene flow leaves a distinctive signature in common population genetic summary statistics, potentially a promising method for evaluating the timing of migration events in speciation genomic studies. Together, these results suggest gene flow likely plays an underappreciated role in avian speciation in tropical mountains, highlighting the need to encompass increased complexity in verbal and quantitative models.