Answers are grounded in this paper and cite their sources.
The target review by Ravinet et al. (2017) gives a comprehensive overview on the population genetic principles governing the build up of reproductive isolation. It illustrates the intricacies of a continuous process unfolding in space and time, involving multiple, interacting barriers. This makes it challenging to pinpoint the underlying mechanisms and isolate the molecular basis of genetic elements (barrier loci) promoting the transition from interconnected populations to independent evolutionary lineages. In their review, Ravinet et al. (2017) connect theoretical expectations with empirically accessible genomic signals that can be extracted from genome scans, which constitute the central empirical tool in the field of speciation genomic research. Genomic regions with unusual levels and spectra of genetic diversity (e.g. π, Tajima's D, Fay and Wu's H), differentiation (e.g. FST) or divergence (e.g. Dxy) are of special interests, as they may represent candidates for selection limiting gene flow, and hence contribute to reproductive isolation. In contrast to a scenario where a large number of genetic changes spread over the genome contribute to reproductive isolation, the genic view of speciation assumed here anticipates that loci of particular importance for reducing gene flow leave genomic signatures detectable by genome scans. The first efforts with the ambition to scan large parts of the genome employed either microsatellites (e.g. Mäkinen et al., 2008) or SNP markers (see Harr, 2006 for an early study), and used the upper tail of FST estimates to nominate candidate loci for divergent selection involved with local adaptation or speciation (similarly, the lower tail was often used to nominate candidates for balancing selection). When genome assemblies started to become available for nonmodel organisms and scans went from marker-based to based on whole-genome (or reduced representation) resequencing, outliers began to take the form of peak regions in the genome with elevated differentiation (Turner et al., 2005). Using the analogy of islands (differentiation peaks) rising over the sea level (genomic background), the terms ‘divergence islands’, ‘differentiation islands’ and even ‘speciation islands’ came in use. They were understood, or at least interpreted, as resulting from resistance to gene flow between diverging lineages. After an initial period of excitement, debate arose as to whether differentiation islands actually translate into genomic regions housing barrier loci (or ‘speciation genes’) involved in reproductive isolation (Noor & Bennett, 2009; Turner & Hahn, 2010; Pennisi, 2014). It has become increasingly clear that heterogeneous divergence landscape can be entirely unrelated to speciation and evolve without differences in the amount of gene flow across the genome (Cruickshank & Hahn, 2014). The field is now maturing with the current state-of-the-art summarized by Seehausen et al. (2014) and more recently by Wolf & Ellegren (2017). Ravinet et al. (2017) propose a further road map towards identifying barrier loci, and carefully acknowledge that accounting for a number of genomic features potentially giving rise to similar signals as divergent selection and/or variable local levels of gene flow can be challenging. Considerable attention has recently been paid to this issue, in particular how linked selection is bound to reduce the effective population size (Ne) in regions of low recombination and high density of functional sites, thereby elevating differentiation via genetic drift (Cutter & Payseur, 2013). Ravinet et al. (2017) emphasize that it remains difficult to empirically establish the causing processes by means of genome scans. It has been rightly proposed that population assemblages rather than isolated population pairs (‘speciation continuum’) may help separate signatures of selection implicated in reducing gene flow from those resulting from variance introduced by demographic change or other forms of linked selection unrelated to speciation (Seehausen et al., 2014; Wolf & Ellegren, 2017). However, genomic signatures of such young evolutionary lineages are nonindependent due to shared recent common ancestry impeding inference. It may therefore be worth considering the possibility of macro-level approaches extending population comparisons beyond closely related taxa to phylogenetically distant clades where lineage sorting has long been completed. Under the assumption that speciation is a context-dependent process involving specific barrier loci for each population pair, covariation in the landscape of genetic diversity across phylogenetically divergent clades is not expected. Genomic parallelism in patterns of genetic diversity (here defined as similar genomic patterns arising in different lineages after lineage sorting is completed and hence being the result of independent evolutionary trajectories) across large evolutionary distances would instead provide evidence for a common process operating in syntenic regions mediated by shared, intrinsic genomic properties. Hence, the comparison of population-specific patterns within (potentially related to speciation) and among clades (unrelated to speciation) in principle provides information on the relative contribution of recent population-specific and common, continuous processes shared across phylogenetic timescales. This macro-level approach has recently been successfully applied in some avian speciation models as described below. Burri et al. (2015) observed parallel genomic landscapes of diversity, differentiation and divergence within and among populations of several flycatcher species within the genus Ficedula. Genomewide variation was associated with gene density and recombination rate derived from one of the species. The results were consistent with a model of linked selection, in particular background selection. The results were further corraborated by the observation of a conserved landscape of recombination rate variation among Ficedula flycatchers (Kawakami et al., 2017). Poelstra et al. (2014) and Vijay et al. (2016) likewise observed covariation in genetic diversity, differentiation, divergence and population recombination rate among populations within and between species in the genus Corvus. They similarly concluded that linked selection common to all populations and unrelated to divergent selection promoting reproductive isolation substantially contributes to the observed heterogeneity. In both systems, flycatchers and crows, elevated levels of differentiation not only seem to be associated with low levels of recombination, but also with structural chromosomal features such as candidate peri-centromeric and subtelomeric regions (Ellegren et al., 2012; Weissensteiner et al., 2017). Van Doren et al. (2017) expanded the phylogenetic framework to comparisons between flycatchers and stonechats, which are different genera residing within the same family of Muscicapida. Similar to the comparisons within genera mentioned above, they found evidence for covariation in FST, Dxy and π and concluded that linked selection has continuously eroded genetic diversity in syntenic genomic regions on each of the branches leading to current day taxa. Expanding the phylogenetic distance even further, Dutoit et al. (2017) asked the question whether local levels of nucleotide diversity were correlated in syntenic regions between collared flycatcher (Ficedula albicollis) and hooded crow (Corvus cornix), two species with an estimated divergence of at least 25 million years. After such long time (4–12 million generations), no polymorphisms present in a common ancestor are expected to remain, and patterns of genetic diversity seen in the contemporary species are thus the result of lineage-specific mutations and evolutionary processes independently affecting them in the two lineages. Using extensive genomic resequencing data, they found π to be well correlated between species. This was not simply due to varying density of coding sequences (which are generally less polymorphic than introns and intergenic DNA), potentially similar in the two lineages, as the correlation was about equally strong when coding sequences were masked. Heterogeneity in levels of genetic diversity was best explained by GC content, pedigree-based estimates of recombination rate and gene density. Mutation rate approximated by the substitution rate of four-fold degenerate sites showed a significant, yet smaller contribution. Given π = 4Neμ, these results are consistent with predictions of linked selection. Vijay et al. (2017) recently extended the crow-flycatcher comparison by adding sequence data from additional species and populations from each system and including a third lineage, Darwin's finches. Using a total of 444 resequenced genomes, they investigated covariation in a number of summary statistics reflecting genomewide variation in Ne across all phylogenetic levels ranging from several thousand to a minimum of 25 million years. They found that not only π but also FST, Dxy and ρ were significantly correlated across all evolutionary distances. Moreover, regions with elevated levels of genetic differentiation were also associated with inferred peri-centromeric and subtelomeric regions. These results provide further evidence that linked selection leads to genomewide heterogeneity in genetic diversity and that it does so in similar ways even among distantly related lineages. The comparative approach described in the avian studies cited above powerfully demonstrates that common forces affected heterogeneity in genetic diversity at syntenic regions of the genome. Birds are known for having stable karyotypes, with few interchromosomal rearrangements, possibly due to a low repeat content of avian genomes (Ellegren, 2010). This may promote stability also in the landscape of genomic features that affect linked selection, that is recombination rate and density of targets of selection. Similarly, the landscape of mutation rate variation may remain relatively stable. To test the generality of these observations, we need more taxa with information on both the variation in diversity and differentiation in parallel lineages and the profiles of candidate genomic features. This will help to understand the relative contribution of ‘common’ forces vs. population-specific divergent selection in generating the island type of heterogeneous divergence landscapes.
No takes yet. Share an insight, caveat, or question.
Ellegren et al. (2017) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: