Speciation is one of the most important processes in biology, yet the study of the genomic changes underlying this process is in its infancy. North American warbler species Setophaga townsendi and Setophaga occidentalis hybridize in a stable hybrid zone, following a period of geographic separation. Genomic differentiation accumulated during geographic isolation can be homogenized by introgression at secondary contact, whereas genetic regions that cause low hybrid fitness can be shielded from such introgression. Here, we examined the genomic underpinning of speciation by investigating (1) the genetic basis of divergent pigmentation traits between species, (2) variation in differentiation across the genome, and (3) the evidence for selection maintaining differentiation in the pigmentation genes. Using tens of thousands of single nucleotide polymorphisms (SNPs) genotyped in hundreds of individuals within and near the hybrid zone, genome-wide association mapping revealed a single SNP associated with cheek, crown, breast coloration, and flank streaking, reflecting pleiotropy (one gene affecting multiple traits) or close physical linkage of different genes affecting different traits. This SNP is within an intron of the RALY gene, hence we refer to it as the RALY SNP. We then examined between-species genomic differentiation, using both genotyping-by-sequencing and whole genome sequencing. We found that the RALY SNP is within one of the highest peaks of differentiation, which contains three genes known to influence pigmentation: ASIP, EIF2S2, and RALY (the ASIP-RALY gene block). Heterozygotes at this gene block are likely of reduced fitness, as the geographic cline of the RALY SNP has been narrow over two decades. Together, these results reflect at least one barrier to gene flow within this narrow (∼200 kb) genomic region that modulates plumage difference between species. Despite extensive gene flow between species across the genome, this study provides evidence that selection on a phenotype-associated genomic region maintains a stable species boundary.

21 22 23 45 RALY gene block maintains stable genotypic and plumage difference between species allowing 46 further differentiation to accumulate via linkage to its flanking genetic region or linkage-47 disequilibrium genome-wide. 48 49 50 extent of yellow plumage) and back colors (green plumage in the mantle), Rohwer & Wood [43] 99predicted that face coloration of Hermit Warblers and hybrids would be controlled by a single-100 locus dominant allele. In addition, whether the other carotenoid and melanin patterning 101 differences between species are underpinned by the same genetic mechanism is an open question. 102These signals can be important in male-male competition for territories, as individuals in the 103 hybrid zone demonstrate an aggression bias towards different plumage types [42,45]. 104We investigated whether speciation in this young species pair occurred through gradual 105 widespread differentiation across the genome or via a small number of restricted regions. If 106 selection for coloration phenotypes has targeted a small number of specific genomic regions, 107 genomic islands of differentiation should be narrow, dispersed, and associated with loci involved 108 in pigmentation pathways [3,7,8]. In addition, we asked the following questions regarding the 109 genetic basis of species-specific traits: (1) is there an island of divergence pleiotropically 110 underpins species-specific features (crown and cheek darkening, breast yellow, bib size, greenish 111 back); (2) Is cheek darkening influenced largely by an allele of dominant effect, consistent with 112 the prediction made two decades ago by Rohwer & Wood (1998); and (3) Is there stable selection 113 maintaining differentiation in the gene region underlying key species diagnostic features? 114 115 Materials and Methods 116 Sampling 117 Whole specimens of Setophaga warblers were collected over two historical sampling 118 sessions in the Washington Cascade hybrid zone [41,43]. The first sampling (N = 314 individuals; 119 35 sites) was carried out in 1987-1994 [43], while the second (N = 127; 11 sites) was in 2005-120 2008 [41] and covered a subset of the sites from the original sampling. We accessed these 121 6 specimens at the Burke Museum of Natural History and Culture (University of Washington, 122Seattle, Washington). 123We carried out a third round of sampling using a catch-and-release approach during the 124 breeding season (early May to mid-July) in 2015-16 ( Figure 1B). Upon locating a territorial male 125 by song, a mist net with a playback (of a locally recorded song) at the bottom was set up nearby. 126After capturing an adult, photographs and a blood sample were taken for further analysis. We re-127 sampled the sites that were sampled in 1986-94 [43]. For details, see previous work in this system 128 [46]. 129 Plumage measurements 130Melanin-and carotenoid-based plumage traits allow identification of the two species 131 ( Figure 1B), but there is also variation within each species [43,47]. To quantify plumage variation 132 within and betwe...

The divergence of plumage color genes contributes to songbird raditation. However, the mechanisms by which color gene divergence counteracts gene flow to maintain reproductive isolation during the formation of new species boundaries remain elusive. The hybrid zone between Setophaga occidentalis (SOCC) and S. townsendi (STOW) in the Cascade Range provides a natural observatory to investigate potential behavioral mechanisms underlying divergent selection on color genes. Recently, we found that selection within a single gene block associated with plumage color variation has maintained a stable and narrow hybrid zone. Here, we investigated the potential role of plumage signals in moderating a behavioral mechanism of selection. Specifically, we assessed whether two plumage traits are associated with body size among breeding males and if trait mismatch predicted aggressive behavior in response to simulated territorial intrusion within hybrid and parental individuals. The two plumage signals, cheek and flank coloration, though associated with the same gene block, reflect opposing dominance of SOCC and STOW alleles. We found that both plumage traits significantly predict the body size in the territorial sex (i.e. males). The opposing dominance of the single color gene block resulted in plumage signal discordance in heterozygotes, which in turn was associated with reduced hybrid territorial performance, an important proxy of fitness in this system. Taken together, these observations point to a single-locus-two-alleles mechanism of incompatibility in shaping a natural species boundary in the early stage of speciation.

Color divergence is increasingly recognized as important for speciation in songbirds through its influence on social dynamics. However, the behavioral mechanisms underlying the eco-evolutionary feedback that acts across species boundaries is poorly understood. The hybrid zone between Setophaga occidentalis (SOCC) and S. townsendi (STOW) in the Cascade mountain ranges provides a natural observatory to test the interplay between genetics and social behaviour in maintaining species boundaries. Recently, we found that selection within a gene block underpinning color variation (ASIP-RALY) has maintained a stable and narrow hybrid zone. Here we investigated the social signaling roles of cheek darkness and flank streaking, two color traits linked to ASIP-RALY that reflect opposing dominance of SOCC and STOW alleles. We found that both traits act as honest badges of status, as they predicted male breeding quality. The opposing dominance effects of ASIP-RALY resulted in signal discordance for heterozygotes, which in turn was associated with inferior hybrid territorial performance, a fitness proxy quantified by vocal and physical responses of resident males to a decoy intruder. Taken together, this study highlights a potential behavioral mechanism underlying selection acting on a simple genetic architecture that has maintained a stable species boundary over decades despite significant gene flow.