Evidence of Site-Specific and Male-Biased Germline Mutation Rate in a Wild Songbird.

Germline mutations are the ultimate source of genetic variation and the raw material for organismal evolution. Despite their significance, the frequency and genomic locations of mutations, as well as potential sex bias, are yet to be widely investigated in most species. To address these gaps, we conducted whole-genome sequencing of 12 great reed warblers (Acrocephalus arundinaceus) in a pedigree spanning 3 generations to identify single-nucleotide de novo mutations (DNMs) and estimate the germline mutation rate. We detected 82 DNMs within the pedigree, primarily enriched at CpG sites but otherwise randomly located along the chromosomes. Furthermore, we observed a pronounced sex bias in DNM occurrence, with male warblers exhibiting three times more mutations than females. After correction for false negatives and adjusting for callable sites, we obtained a mutation rate of 7.16 × 10-9 mutations per site per generation (m/s/g) for the autosomes and 5.10 × 10-9 m/s/g for the Z chromosome. To demonstrate the utility of species-specific mutation rates, we applied our autosomal mutation rate in models reconstructing the demographic history of the great reed warbler. We uncovered signs of drastic population size reductions predating the last glacial period (LGP) and reduced gene flow between western and eastern populations during the LGP. In conclusion, our results provide one of the few direct estimates of the mutation rate in wild songbirds and evidence for male-driven mutations in accordance with theoretical expectations.

Diurnal migration patterns in willow warblers differ between the western and eastern flyways.

It is a long-standing view that the main mechanism maintaining narrow migratory divides in passerines is the selection against intermediate and suboptimal migratory direction, but empirical proof of this is still lacking. We present novel results from a willow warbler migratory divide in central Sweden from where birds take the typical SW and SE as well as intermediate routes to winter quarters in Africa. We hypothesized that individuals that take the intermediate route are forced to migrate in daytime more often when crossing wide ecological barriers than birds that follow the typical western or eastern flyways. Analyses of geolocator tracks of willow warblers breeding across the entire Sweden, including the migratory divide, provided no support for our hypothesis. Instead, birds that migrated along the western flyway were the most likely to undertake full day flights. The probability of migrating for a full day when crossing major barriers declined linearly from west to east. We speculate that this difference is possibly caused by more challenging conditions in the western part of the Sahara Desert, such as the lack of suitable day-time roost sites. However, it may equally likely be that willow warblers benefit from migrating in daytime if favorable tailwinds offer assistance.

Improved haplotype resolution of highly duplicated MHC genes in a long-read genome assembly using MiSeq amplicons.

Long-read sequencing offers a great improvement in the assembly of complex genomic regions, such as the major histocompatibility complex (MHC) region, which can contain both tandemly duplicated MHC genes (paralogs) and high repeat content. The MHC genes have expanded in passerine birds, resulting in numerous MHC paralogs, with relatively high sequence similarity, making the assembly of the MHC region challenging even with long-read sequencing. In addition, MHC genes show rather high sequence divergence between alleles, making diploid-aware assemblers incorrectly classify haplotypes from the same locus as sequences originating from different genomic regions. Consequently, the number of MHC paralogs can easily be over- or underestimated in long-read assemblies. We therefore set out to verify the MHC diversity in an original and a haplotype-purged long-read assembly of one great reed warbler Acrocephalus arundinaceus individual (the focal individual) by using Illumina MiSeq amplicon sequencing. Single exons, representing MHC class I (MHC-I) and class IIB (MHC-IIB) alleles, were sequenced in the focal individual and mapped to the annotated MHC alleles in the original long-read genome assembly. Eighty-four percent of the annotated MHC-I alleles in the original long-read genome assembly were detected using 55% of the amplicon alleles and likewise, 78% of the annotated MHC-IIB alleles were detected using 61% of the amplicon alleles, indicating an incomplete annotation of MHC genes. In the haploid genome assembly, each MHC-IIB gene should be represented by one allele. The parental origin of the MHC-IIB amplicon alleles in the focal individual was determined by sequencing MHC-IIB in its parents. Two of five larger scaffolds, containing 6-19 MHC-IIB paralogs, had a maternal and paternal origin, respectively, as well as a high nucleotide similarity, which suggests that these scaffolds had been incorrectly assigned as belonging to different loci in the genome rather than as alternate haplotypes of the same locus. Therefore, the number of MHC-IIB paralogs was overestimated in the haploid genome assembly. Based on our findings we propose amplicon sequencing as a suitable complement to long-read sequencing for independent validation of the number of paralogs in general and for haplotype inference in multigene families in particular.

Balancing selection on the complement system of a wild rodent.

BACKGROUND: Selection pressure exerted by pathogens can influence patterns of genetic diversity in the host. In the immune system especially, numerous genes encode proteins involved in antagonistic interactions with pathogens, paving the way for coevolution that results in increased genetic diversity as a consequence of balancing selection. The complement system is a key component of innate immunity. Many complement proteins interact directly with pathogens, either by recognising pathogen molecules for complement activation, or by serving as targets of pathogen immune evasion mechanisms. Complement genes can therefore be expected to be important targets of pathogen-mediated balancing selection, but analyses of such selection on this part of the immune system have been limited. RESULTS: Using a population sample of whole-genome resequencing data from wild bank voles (n = 31), we estimated the extent of genetic diversity and tested for signatures of balancing selection in multiple complement genes (n = 44). Complement genes showed higher values of standardised ß (a statistic expected to be high under balancing selection) than the genome-wide average of protein coding genes. One complement gene, FCNA, a pattern recognition molecule that interacts directly with pathogens, was found to have a signature of balancing selection, as indicated by the Hudson-Kreitman-Aguadé test (HKA) test. Scans for localised signatures of balancing selection in this gene indicated that the target of balancing selection was found in exonic regions involved in ligand binding. CONCLUSION: The present study adds to the growing evidence that balancing selection may be an important evolutionary force on components of the innate immune system. The identified target in the complement system typifies the expectation that balancing selection acts on genes encoding proteins involved in direct interactions with pathogens.

Migration direction in a songbird explained by two loci.

Migratory routes and remote wintering quarters in birds are often species and even population specific. It has been known for decades that songbirds mainly migrate solitarily, and that the migration direction is genetically controlled. Yet, the underlying genetic mechanisms remain unknown. To investigate the genetic basis of migration direction, we track genotyped willow warblers Phylloscopus trochilus from a migratory divide in Sweden, where South-West migrating, and South-East migrating subspecies form a hybrid swarm. We find evidence that migration direction follows a dominant inheritance pattern with epistatic interaction between two loci explaining 74% of variation. Consequently, most hybrids migrate similarly to one of the parental subspecies, and therefore do not suffer from the cost of following an inferior, intermediate route. This has significant implications for understanding the selection processes that maintain narrow migratory divides.

Inversions maintain differences between migratory phenotypes of a songbird.

Structural rearrangements have been shown to be important in local adaptation and speciation, but have been difficult to reliably identify and characterize in non-model species. Here we combine long reads, linked reads and optical mapping to characterize three divergent chromosome regions in the willow warbler Phylloscopus trochilus, of which two are associated with differences in migration and one with an environmental gradient. We show that there are inversions (0.4-13 Mb) in each of the regions and that the divergence times between inverted and non-inverted haplotypes are similar across the regions (~1.2 Myrs), which is compatible with a scenario where inversions arose in either of two allopatric populations that subsequently hybridized. The improved genomes allow us to detect additional functional differences in the divergent regions, providing candidate genes for migration and adaptations to environmental gradients.

Extreme variation in recombination rate and genetic diversity along the Sylvioidea neo-sex chromosome.

Recombination strongly impacts sequence evolution by affecting the extent of linkage and the efficiency of selection. Here, we study recombination over the Z chromosome in great reed warblers (Acrocephalus arundinaceus) using pedigree-based linkage mapping. This species has extended Z and W chromosomes ("neo-sex chromosomes") formed by a fusion between a part of chromosome 4A and the ancestral sex chromosomes, which provides a unique opportunity to assess recombination and sequence evolution in sex-linked regions of different ages. We assembled an 87.54 Mbp and 90.19 cM large Z with a small pseudoautosomal region (0.89 Mbp) at one end and the fused Chr4A-part at the other end of the chromosome. A prominent feature in our data was an extreme variation in male recombination rate along Z with high values at both chromosome ends, but an apparent lack of recombination over a substantial central section, covering 78% of the chromosome. The nonrecombining region showed a drastic loss of genetic diversity and accumulation of repeats compared to the recombining parts. Thus, our data emphasize a key role of recombination in affecting local levels of polymorphism. Nonetheless, the evolutionary rate of genes (dN/dS) did not differ between high and low recombining regions, suggesting that the efficiency of selection on protein-coding sequences can be maintained also at very low levels of recombination. Finally, the Chr4A-derived part showed a similar recombination rate as the part of the ancestral Z that did recombine, but its sequence characteristics reflected both its previous autosomal, and current Z-linked, recombination patterns.

The genomic architecture of the passerine MHC region: High repeat content and contrasting evolutionary histories of single copy and tandemly duplicated MHC genes.

The major histocompatibility complex (MHC) is of central importance to the immune system, and an optimal MHC diversity is believed to maximize pathogen elimination. Birds show substantial variation in MHC diversity, ranging from few genes in most bird orders to very many genes in passerines. Our understanding of the evolutionary trajectories of the MHC in passerines is hampered by lack of data on genomic organization. Therefore, we assembled and annotated the MHC genomic region of the great reed warbler (Acrocephalus arundinaceus), using long-read sequencing and optical mapping. The MHC region is large (>5.5 Mb), characterized by structural changes compared to hitherto investigated bird orders and shows higher repeat content than the genome average. These features were supported by analyses in three additional passerines. MHC genes in passerines are found in two different chromosomal arrangements, either as single copy MHC genes located among non-MHC genes, or as tandemly duplicated tightly linked MHC genes. Some single copy MHC genes are old and putative orthologues among species. In contrast tandemly duplicated MHC genes are monophyletic within species and have evolved by simultaneous gene duplication of several MHC genes. Structural differences in the MHC genomic region among bird orders seem substantial compared to mammals and have possibly been fuelled by clade-specific immune system adaptations. Our study provides methodological guidance in characterizing complex genomic regions, constitutes a resource for MHC research in birds, and calls for a revision of the general belief that avian MHC has a conserved gene order and small size compared to mammals.

Transposable elements mark a repeat-rich region associated with migratory phenotypes of willow warblers (Phylloscopus trochilus).

The genetic basis of bird migration has been the focus of several studies. Two willow warbler subspecies (Phylloscopus trochilus trochilus and Phylloscopus trochilus acredula) follow different migratory routes to wintering grounds in Africa. Their breeding populations overlap in contact areas or "migratory divides" located in central Scandinavia and in eastern Poland. Earlier analyses demonstrated that the genetic differences between these two migratory phenotypes are few and cluster on chromosomes 1 and 5. In addition, an amplified fragment length polymorphism-derived biallelic marker (known as WW2) presents steep clines across both migratory divides but failed to be mapped in the genome. Here, we characterize the WW2 marker and describe its two variants (WW2 ancestral and WW2 derived) as portions of long terminal repeat retrotransposons originating from an ancient infection by an endogenous retrovirus. We used quantitative polymerase chain reaction techniques to quantify copy numbers of the WW2 derived variant in the two subspecies and their hybrids. This, together with genome analyses revealed that WW2 derived variants are much more abundant in P. t. acredula and appear embedded in a large repeat-rich region (>12 Mbp), not associated with the divergent regions of chromosomes 1 or 5. However, it might interact with genetic elements controlling migration direction. Testing this hypothesis further will require knowing the exact location of this region, such as by obtaining more complete genome assemblies preferably in combination with techniques like fluorescence in situ hybridization applied to a willow warbler karyotype, and finally to investigate the copy number of this marker in hybrids with known migratory tracks.

Avian Neo-Sex Chromosomes Reveal Dynamics of Recombination Suppression and W Degeneration.

How the avian sex chromosomes first evolved from autosomes remains elusive as 100 million years (My) of divergence and degeneration obscure their evolutionary history. The Sylvioidea group of songbirds is interesting for understanding avian sex chromosome evolution because a chromosome fusion event ∼24 Ma formed "neo-sex chromosomes" consisting of an added (new) and an ancestral (old) part. Here, we report the complete female genome (ZW) of one Sylvioidea species, the great reed warbler (Acrocephalus arundinaceus). Our long-read assembly shows that the added region has been translocated to both Z and W, and whereas the added-Z has retained its gene order the added-W part has been heavily rearranged. Phylogenetic analyses show that recombination between the homologous added-Z and -W regions continued after the fusion event, and that recombination suppression across this region took several million years to be completed. Moreover, recombination suppression was initiated across multiple positions over the added-Z, which is not consistent with a simple linear progression starting from the fusion point. As expected following recombination suppression, the added-W show signs of degeneration including repeat accumulation and gene loss. Finally, we present evidence for nonrandom maintenance of slowly evolving and dosage-sensitive genes on both ancestral- and added-W, a process causing correlated evolution among orthologous genes across broad taxonomic groups, regardless of sex linkage.

Continent-wide genomic signatures of adaptation to urbanisation in a songbird across Europe.

Urbanisation is increasing worldwide, and there is now ample evidence of phenotypic changes in wild organisms in response to this novel environment. Yet, the genetic changes and genomic architecture underlying these adaptations are poorly understood. Here, we genotype 192 great tits (Parus major) from nine European cities, each paired with an adjacent rural site, to address this major knowledge gap in our understanding of wildlife urban adaptation. We find that a combination of polygenic allele frequency shifts and recurrent selective sweeps are associated with the adaptation of great tits to urban environments. While haplotypes under selection are rarely shared across urban populations, selective sweeps occur within the same genes, mostly linked to neural function and development. Collectively, we show that urban adaptation in a widespread songbird occurs through unique and shared selective sweeps in a core-set of behaviour-linked genes.

Divergence in Coding Sequence and Expression of Different Functional Categories of Immune Genes between Two Wild Rodent Species.

Differences in immune function between species could be a result of interspecific divergence in coding sequence and/or expression of immune genes. Here, we investigate how the degree of divergence in coding sequence and expression differs between functional categories of immune genes, and if differences between categories occur independently of other factors (expression level, pleiotropy). To this end, we compared spleen transcriptomes of wild-caught yellow-necked mice and bank voles. Immune genes expressed in the spleen were divided into four categories depending on the function of the encoded protein: pattern recognition receptors (PRR); signal transduction proteins; transcription factors; and cyto- and chemokines and their receptors. Genes encoding PRR and cyto-/chemokines had higher sequence divergence than genes encoding signal transduction proteins and transcription factors, even when controlling for potentially confounding factors. Genes encoding PRR also had higher expression divergence than genes encoding signal transduction proteins and transcription factors. There was a positive correlation between expression divergence and coding sequence divergence, in particular for PRR genes. We propose that this is a result of that divergence in PRR coding sequence leads to divergence in PRR expression through positive feedback of PRR ligand binding on PRR expression. When controlling for sequence divergence, expression divergence of PRR genes did not differ from other categories. Taken together, the results indicate that coding sequence divergence of PRR genes is a major cause of differences in immune function between species.

Comparison of spleen transcriptomes of two wild rodent species reveals differences in the immune response against Borrelia afzelii.

Different host species often differ considerably in susceptibility to a given pathogen, but the causes of such differences are rarely known. The natural hosts of the tick-transmitted bacterium Borrelia afzelii, which is one of causative agents of Lyme borreliosis in humans, include a variety of small mammals like voles and mice. Previous studies have shown that B. afzelii-infected bank voles (Myodes glareolus) have about ten times higher bacterial load than infected yellow-necked mice (Apodemus flavicollis), indicating that these two species differ in resistance. In this study, we compared the immune response to B. afzelii infection in these host species by using RNA sequencing to quantify gene expression in spleen. Gene set enrichment analysis (GSEA) showed that several immune pathways were down-regulated in infected animals in both bank voles and yellow-necked mice. Moreover, IFNα response was up-regulated in B. afzelii-infected yellow-necked mice, while IL6 signaling and the complement pathway were down-regulated in infected bank voles; differences in regulation of these three pathways between bank voles and yellow-necked mice could thus contribute to the difference in resistance to B. afzelii between the species. This study provides knowledge of gene expression induced by a zoonotic pathogen in its natural host, and possible species-specific regulation of immune responses associated with resistance.

Autumn migration direction of juvenile willow warblers (Phylloscopus t. trochilus and P. t. acredula) and their hybrids assessed by qPCR SNP genotyping.

BACKGROUNDS: Geographic regions, where two closely related taxa with different migration routes come into contact, are known as migratory divides. Hybrids originating from migratory divides are hypothesized to migrate intermediately relative to the parental populations. Few studies have tested this hypothesis in wild birds, and only in hybrids that have completed the migration back to the breeding grounds. Here, we make use of the well-established migration routes of willow warblers (Phylloscopus trochilus), for which the subspecies trochilus and acredula have migration-associated genetic markers on chromosomes 1 and 5. The genetic approach enabled us to analyze the geographic distribution of juveniles during their first autumn migration, predicting that hybrids should be more frequent in the central flyway over Italy than along the typical SW routes of trochilus and SE routes of acredula. METHODS: Blood and feather samples were collected from wintering birds in Africa (n = 69), and from juveniles during autumn migration in Portugal (n = 33), Italy (n = 38) and Bulgaria (n = 32). Genotyping was carried out by qPCR SNP assays, on one SNP each on chromosome 1 (SNP 65) and chromosome 5 (SNP 285). Both these SNPs have alternative alleles that are highly fixed (> 97%) in each of the subspecies. RESULTS: The observed combined genotypes of the two SNPs were associated with the known migration routes and wintering distributions of trochilus and acredula, respectively. We found hybrids (HH) among the juveniles in Italy (5/38) and in Portugal (2/33). The proportion of hybrids in Italy was significantly higher than expected from a background rate of hybrid genotypes (1.5%) in allopatric populations of the subspecies. CONCLUSIONS: Our genetic approach to assign individuals to subspecies and hybrids allowed us to investigate migration direction in juvenile birds on their first migration, which should better reflect the innate migratory direction than studies restricted to successful migrants. The excess of hybrids in Italy, suggests that they employ an intermediate route relative to the parental populations. Our qPCR SNP genotyping method is efficient for processing large sample sizes, and will therefore be useful in migration research of species with known population genetic structure.

Balancing selection in Pattern Recognition Receptor signalling pathways is associated with gene function and pleiotropy in a wild rodent.

Pathogen-mediated balancing selection is commonly considered to play an important role in the maintenance of genetic diversity, in particular in immune genes. However, the factors that may influence which immune genes are the targets of such selection are largely unknown. To address this, here we focus on Pattern Recognition Receptor (PRR) signalling pathways, which play a key role in innate immunity. We used whole-genome resequencing data from a population of bank voles (Myodes glareolus) to test for associations between balancing selection, pleiotropy and gene function in a set of 123 PRR signalling pathway genes. To investigate the effect of gene function, we compared genes encoding (a) receptors for microbial ligands versus downstream signalling proteins, and (b) receptors recognizing components of microbial cell walls, flagella and capsids versus receptors recognizing features of microbial nucleic acids. Analyses based on the nucleotide diversity of full coding sequences showed that balancing selection primarily targeted receptor genes with a low degree of pleiotropy. Moreover, genes encoding receptors recognizing components of microbial cell walls etc. were more important targets of balancing selection than receptors recognizing nucleic acids. Tests for localized signatures of balancing selection in coding and noncoding sequences showed that such signatures were mostly located in introns, and more evenly distributed among different functional categories of PRR pathway genes. The finding that signatures of balancing selection in full coding sequences primarily occur in receptor genes, in particular those encoding receptors for components of microbial cell walls etc., is consistent with the idea that coevolution between hosts and pathogens is an important cause of balancing selection on immune genes.

Comparative analysis examining patterns of genomic differentiation across multiple episodes of population divergence in birds.

Heterogeneous patterns of genomic differentiation are commonly documented between closely related populations and there is considerable interest in identifying factors that contribute to their formation. These factors could include genomic features (e.g., areas of low recombination) that promote processes like linked selection (positive or purifying selection that affects linked neutral sites) at specific genomic regions. Examinations of repeatable patterns of differentiation across population pairs can provide insight into the role of these factors. Birds are well suited for this work, as genome structure is conserved across this group. Accordingly, we reestimated relative (FST ) and absolute (dXY ) differentiation between eight sister pairs of birds that span a broad taxonomic range using a common pipeline. Across pairs, there were modest but significant correlations in window-based estimates of differentiation (up to 3% of variation explained for FST and 26% for dXY ), supporting a role for processes at conserved genomic features in generating heterogeneous patterns of differentiation; processes specific to each episode of population divergence likely explain the remaining variation. The role genomic features play was reinforced by linear models identifying several genomic variables (e.g., gene densities) as significant predictors of FST and dXY repeatability. FST repeatability was higher among pairs that were further along the speciation continuum (i.e., more reproductively isolated) providing further insight into how genomic differentiation changes with population divergence; early stages of speciation may be dominated by positive selection that is different between pairs but becomes integrated with processes acting according to shared genomic features as speciation proceeds.

Genetic differences between willow warbler migratory phenotypes are few and cluster in large haplotype blocks.

It is well established that differences in migratory behavior between populations of songbirds have a genetic basis but the actual genes underlying these traits remains largely unknown. In an attempt to identify such candidate genes we de novo assembled the genome of the willow warbler Phylloscopus trochilus, and used whole-genome resequencing and a SNP array to associate genomic variation with migratory phenotypes across two migratory divides around the Baltic Sea that separate SW migrating P. t. trochilus wintering in western Africa and SSE migrating P. t. acredula wintering in eastern and southern Africa. We found that the genomes of the two migratory phenotypes lack clear differences except for three highly differentiated regions located on chromosomes 1, 3, and 5 (containing 146, 135, and 53 genes, respectively). Within each migratory phenotype we found virtually no differences in allele frequencies for thousands of SNPs, even when comparing geographically distant populations breeding in Scandinavia and Far East Russia (>6000 km). In each of the three differentiated regions, multidimensional scaling-based clustering of SNP genotypes from more than 1100 individuals demonstrates the presence of distinct haplotype clusters that are associated with each migratory phenotype. In turn, this suggests that recombination is absent or rare between haplotypes, which could be explained by inversion polymorphisms. Whereas SNP alleles on chromosome 3 correlate with breeding altitude and latitude, the allele distribution within the regions on chromosomes 1 and 5 perfectly matches the geographical distribution of the migratory phenotypes. The most differentiated 10 kb windows and missense mutations within these differentiated regions are associated with genes involved in fatty acid synthesis, possibly representing physiological adaptations to the different migratory strategies. The ∼200 genes in these regions, of which several lack described function, will direct future experimental and comparative studies in the search for genes that underlie important migratory traits.

Gene expression in the brain of a migratory songbird during breeding and migration.

BACKGROUND: We still have limited knowledge about the underlying genetic mechanisms that enable migrating species of birds to navigate the globe. Here we make an attempt to get insight into the genetic architecture controlling this complex innate behaviour. We contrast the gene expression profiles of two closely related songbird subspecies with divergent migratory phenotypes. In addition to comparing differences in migratory strategy we include a temporal component and contrast patterns between breeding adults and autumn migrating juvenile birds of both subspecies. The two willow warbler subspecies, Phylloscopus trochilus trochilus and P. t. acredula, are remarkably similar both in phenotype and genotype and have a narrow contact zone in central Scandinavia. Here we used a microarray gene chip representing 23,136 expressed sequence tags (ESTs) from the zebra finch Taeniopygia guttata to identify mRNA level differences in willow warbler brain tissue in relation to subspecies and season. RESULTS: Out of the 22,109 EST probe sets that remained after filtering poorly binding probes, we found 11,898 (51.8 %) probe sets that could be reliably and uniquely matched to a total of 6,758 orthologous zebra finch genes. The two subspecies showed very similar levels of gene expression with less than 0.1 % of the probe sets being significantly differentially expressed. In contrast, 3,045 (13.8 %) probe sets were found to be differently regulated between samples collected from breeding adults and autumn migrating juvenile birds. The genes found to be differentially expressed between seasons appeared to be enriched for functional roles in neuronal firing and neuronal synapse formation. CONCLUSIONS: Our results show that only few genes are differentially expressed between the subspecies. This suggests that the different migration strategies of the subspecies might be governed by few genes, or that the expression patterns of those genes are time-structured or tissue-specific in ways, which our approach fails to uncover. Our findings will be useful in the planning of new experiments designed to unravel the genes involved in the migratory program of birds.

No evidence for assortative mating within a willow warbler migratory divide.

INTRODUCTION: In contact zones, genetic mixing of two taxa can be restricted by prezygotic (e.g. assortative mating) or postzygotic (lower fitness of hybrid offspring) barriers, or a combination of the two. A hybrid zone between two willow warbler subspecies (Phylloscopus trochilus trochilus, P. t. acredula) with distinctive migratory strategies occurs in central Sweden. These subspecies exhibit differences in migratory direction and distance, resulting in geographically distinct wintering areas in Africa. The subspecies may have diverged from a common refuge after the last ice age, and neutral genetic markers are homogeneous across their range. By contrast, several phenotypic traits and genetic markers of two chromosomal regions previously identified show steep clines across the divide. The evolutionary forces that maintain this migratory divide remain unknown. Here we use plumage colour, morphology, genetic markers and feather stable nitrogen-isotopes (δ (15)N) to assess if assortative mating between migratory phenotypes could be acting as a possible mechanism for keeping the two forms genetically separate and maintaining the migratory divide. We colour-ringed a willow warbler breeding population in the central part of the hybrid zone and observed the breeding population to assess phenotypic and genotypic traits of social pairs. RESULTS: Our data suggest that wintering area and genetic ancestry had an effect on male arrival time to the breeding grounds which could contribute to assortment. However, evidence for assortative mating could not be detected based on a comparison of plumage colour, morphology and δ (15)N between social mates. CONCLUSION: This finding was strengthened by analyses of subspecies-specific genetic markers, which allowed us to identify the presence of a large proportion of potential hybrids and backcrosses at the study site. Our results supported the hypothesis that pre-mating isolation in willow warblers is weak, resulting in extensive hybridisation across the migratory divide.

Allelic variation in a willow warbler genomic region is associated with climate clines.

Local adaptation is an important process contributing to population differentiation which can occur in continuous or isolated populations connected by various amounts of gene flow. The willow warbler (Phylloscopus trochilus) is one of the most common songbirds in Fennoscandia. It has a continuous breeding distribution where it is found in all forested habitats from sea level to the tree line and therefore constitutes an ideal species for the study of locally adapted genes associated with environmental gradients. Previous studies in this species identified a genetic marker (AFLP-WW1) that showed a steep north-south cline in central Sweden with one allele associated with coastal lowland habitats and the other with mountainous habitats. It was further demonstrated that this marker is embedded in a highly differentiated chromosome region that spans several megabases. In the present study, we sampled 2,355 individuals at 128 sites across all of Fennoscandia to study the geographic and climatic variables associated with the allele frequency distributions of WW1. Our results demonstrate that 1) allele frequency patterns significantly differ between mountain and lowland populations, 2) these allele differences coincide with extreme temperature conditions and the short growing season in the mountains, and milder conditions in coastal areas, and 3) the northern-allele or "altitude variant" of WW1 occurs in willow warblers that occupy mountainous habitat regardless of subspecies. Finally these results suggest that climate may exert selection on the genomic region associated with these alleles and would allow us to develop testable predictions for the distribution of the genetic marker based on climate change scenarios.