Final Report Summary - NGS-RINGSPS (Searching genomic regions involved in adaptive evolution and speciation using ring species as a model system and next-generation sequencing) Ring species are very valuable model systems in evolutionary biology given that they nicely illustrate how one single species can gradually split into two, new daughter species. Unfortunately, ring species are rare. The greenish warbler Phylloscopus trochiloides is one of the very few well-documented examples in the wild. Greenish warblers form a chain of intergrading populations that encircle the high altitude and treeless desserts of the Tibetan plateau (Ticerhurst, 1938). Interbreeding among neighboring populations has been believed to occur at any point of the ring, except where the two terminal forms of the ring overlap in central Siberia. Greenish warblers have been also suggested to represent a situation where parallel or convergent evolution, rather than divergent evolution, has promoted the reproductive isolation between independent lineages evolving at different sides of the Tibetan plateau. For instance, the length and complexity of song gradually increases from south to the north, northern forms display smaller body sizes and similar habitat preferences and are also prepared to carry out much longer displacements during seasonal migration. Although the structure of song gradually increases, greenish warbler have found different ways to communicate in the more closed forests from the north, where population densities are smaller. As a result, the structure of song greatly differs between the two sympatric forms co-existing in central Siberia, to the point that they have been suggested to be incapable to recognize each other as member of the same species and interbreed. In any case, this evidence for parallel or convergent evolution of independent lineages represents a unique opportunity to figure out whether the same regions of the genome have been responsible for the adaptation to novel conditions as greenish warblers colonized northern areas from a presumably glacial refugee south of the Himalayan chain. Fortunately, the advent of next-generation sequencing technologies permitted us to accomplish such ambitious goals through sophisticated genome-wide scans. Previous to the identification of genome regions showing increased differentiation with respect to the average genome background it was critical to provide a genome-wide update about the genetic relationships among greenish warbler populations. This is so because previous genetic data proved inconclusive with respect to whether the process of species divergence took place in the face of continuous gene flow or whether hybridization between the terminal forms of the ring ever occurred. Such as an update will allow us inferring if natural selection has been the major evolutionary force shaping the genome of greenish warblers or whether periods of geographical isolation (i.e. allopatric divergence) also contributed to and/or accelerated the observed patterns. The analysis of mtDNA sequences (Irwin et al. 2001, Nature) revealed several possible phylogenetic breaks around the ring, the sharpest between the two sympatric forms inhabiting Siberia but also in the western Himalayas. On the contrary, nuclear data based on AFLP (Amplified Fragment Length Polymorphism) data suggested a smooth gradient of genetic variation around the ring, with no apparent breaks and fitting well into an isolation-by-distance model (Irwin et al. 2005, Science). Mitochondrial DNA exhibited a 100% concordance between mtDNA type and bird song, and AFLP showed the two northern forms as clearly unique genetic clusters, with no evidence for recent or past hybridization. During the development of the present project, a genotyping-by-sequencing approach provided the tools (>80,000 Single Nucleotide Polymorphism - SNP markers distributed genome-wide) for a much more in-deep characterization of the genomic relationships among greenish warbler populations and helped us to resolve some of the inconsistencies found between nuclear and mtDNA markers. The analysis of thousand of genetic markers supported data from mtDNA and clearly indicated an ancient period of geographical isolation in the west Himalayas followed by a contemporary period of secondary contact. A detailed analysis across this contact zone revealed the classic features of a hybrid zone between genetically different lineages. Moreover, the extraordinarily increased resolution provided by massive parallel Illumina sequencing of 95 greenish warblers showed compelling evidence of genetic exchange between the two terminal forms. The two Siberian forms overlapping in central Siberia were supposed to have developed complete reproductive isolation by the time they established secondary contact. Our genome-wide analyses were therefore surprising in indicating that up to 10% of the genome of western warblers was introgressed into the genome of eastern warblers. Indeed, these novel findings constituted a tremendous revolution concerning the interpretation of the greenish warbler ring species, one of the most widely examples to teach evolution. Overall, our findings rule out a scenario where a single common ancestor dwelling south of the Himalayan plateau gradually diverged into new species in the face of continuous gene flow. Periods of geographical isolation and secondary contact, as predicted by the constant cyclic climatic upheavals that the Earth experience during the Pleistocene glaciations, should have been common. Therefore, allopatric divergence also might have contributed to shaping the genome of greenish warblers. Secondly, the development of complete reproductive isolation demanded for longer periods of time than those experienced previous to the colonization of the northernmost areas, and rare hybridization has been possible. However, the low levels of asymmetric introgression found in Central Siberia denote strong selective pressures against novel gene combinations in the contact zone but still, hybrids are viable. Either way, the updated picture of the greenish warbler ring species brings into focus two contrasting contact zones. Regardless of the extent of of genome-wide divergence, interbreeding between morphologically and behaviorally similar populations in the south is much more widespread than gene flow among northern forms strongly differing in characters such as song, plumage and migratory behavior. The de-novo discovery and genotyping of more than 80,000 SNPs assisted detailed genome-wide scans. These analyses revealed several peaks of increased genetic differentiation or “islands of divergence” throughout the genome, normally one or two per chromosome. These genomic regions spanned several hundreds of thousands of base pairs and hosted a few dozens of genes. The extent of genetic differentiation within these regions was extraordinarily higher than the average genome background. Fixed nucleotide differences among subspecies were mostly absent from the vast majority of the genome, which denotes the relatively recent evolutionary history of the superspecies complex. Fixed differences among subspecies, on the other hand, cropped quite abundantly within the islands of divergence. Islands of divergence were also characterized by a dramatic decrease in genetic variability and strong linkage disequilibrium. Remarkably, the same regions that diverged most in one side of the ring, with respect to the presumably common ancestor, were also the regions that diverged most in the other side of the ring. Since we believe that the observed patterns are not exclusively caused by the architecture of the genome (i.e. regions of low recombination such as centromeres or telomeres) we believe this study constitutes an outstanding illustration supporting the participation of certain genomic regions during the process of local adaptation, and eventually, the development of reproductive isolation and the emergence of new species. The picture put forward by our genomic scans shows a very heterogeneous landscape with respect to genome differentiation, with only a few regions sticking out from the average genome background. This is certainly what it can be expected under a scenario of divergence with gene flow. Here, neutral regions become permeable to the homogenizing effects of gene flow whilst regions involved in local adaptation develop incompatibilities for genetic exchange. The assembly of the genome of two individuals belonging to the sympatric northern forms (viridanus and plumbeitarsus) and one individual belonging to the presumably common ancestor (trochiloides) allowed us reconstructing the genes contained within the islands of divergence and carried out a very detailed comparison. Here, we found a surprisingly low number of mutations changing the amino acid sequence of the proteins coded by those genes within the core of the island of divergence. This finding therefore suggests a probably major role of regulatory elements in driving local adaptation and speciation, rather than the fast evolution of protein sequences. Furthermore, we also detected an unusually high presence of genes which main function is the regulation of the expression of other genes (e.g. methyltransferases, transcription factors, hormonal receptors or genes that regulate the assembly of the chromatin). Finally, we used the Major Histocompatibility Complex (MHC) as a model system to investigate the spatial distribution of genetic variation in a set of genes strongly subjected to pathogen-mediated pressures. In the same line than traits such as song length or body size, we found that the distribution of allele frequencies at both MHC class I and MHC class II B genes (participating in the adaptive immune response against intracellular and extracellular pathogens, respectively) was more similar between the terminal forms of the ring than across any other population comparison around the ring. These exciting findings supported the hypotheses that natural selection has shaped MHC variation in independent lineages in such a similar way, and as the result of the colonization of similar habitats during the northward expansion that encircled the Tibetan plateau. These findings also constitute the first evidence of parallel or convergent evolution ever supported by molecular data in the greenish warbler ring species.
