How Reticulated Are Species?

How reticulated are species?

The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters

Citation Mallet, James, Nora Besansky, and Matthew W. Hahn. 2015. “How reticulated are species?” Bioessays 38 (2): 140-149. doi:10.1002/ bies.201500149. http://dx.doi.org/10.1002/bies.201500149.

Published Version doi:10.1002/bies.201500149

Citable link http://nrs.harvard.edu/urn-3:HUL.InstRepos:29002679

Terms of Use This article was downloaded from Harvard University’s DASH repository, and is made available under the terms and conditions applicable to Other Posted Material, as set forth at http:// nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of- use#LAA Insights & Perspectives

How reticulated are species? Think again James Mallet1)2)*, Nora Besansky3) and Matthew W. Hahn4)

Many groups of closely related species have reticulate phylogenies. Recent well as microbes. Reproductively iso- genomic analyses are showing this in many insects and vertebrates, as well as lated species and bifurcating phyloge- in microbes and plants. In microbes, lateral gene transfer is the dominant nies have become an important basis for process that spoils strictly tree-like phylogenies, but in multicellular eukaryotes our understanding of evolution; now this bedrock seems threatened. As an hybridization and introgression among related species is probably more ideal, species are often taken to be important. Because many species, including the ancestors of ancient major evolutionarily independent populations lineages, seem to evolve rapidly in adaptive radiations, some sexual that are reproductively isolated from compatibility may exist among them. Introgression and reticulation can thereby other such species, for example in the affect all parts of the tree of life, not just the recent species at the tips. Our “biological species concept,” although it was always known that hybridization understanding of adaptive evolution, speciation, phylogenetics, and compar- does occur [15]. Reticulate evolution in ative biology must adapt to these mostly recent ﬁndings. Introgression has plants has long been recognized [16], important practical implications as well, not least for the management of but recent genomic evidence from genetically modiﬁed organisms in pest and disease control. animals suggest that reticulation might be much more common than antici- Keywords: pated [17, 18]. Given abundant new .admixture; homoplasy; introgression; phylogenetic discordance; speciation; data, it is time to enquire whether a species concepts; tree of life major shift in our understanding of species, speciation, and phylogenetics Introduction tidy tree picture was threatened: is taking place. sequencing of more microbial genes Not so long ago, analysis of microbial and then whole genomes quickly led to 16S ribosomal RNA sequences led to a an understanding of the importance of revolutionary new “Universal Tree of horizontal or lateral gene transfer, the Prokaryotes: Is there a Life,” consisting of three monophyletic incorporation of foreign genes into the universal tree of life? domains, here referred to as the Bacte- genome. Some of the major transitions ria, the Archaea, and the Eukarya or in evolution were clearly due to lateral Tree-like relationships among species eukaryotes [1, 2]. Yet almost as soon as transfer: the eukaryotes were formed by arise because the genome the new system was established, this endosymbiosis of a-proteobacteria with evolves within cells. When a cell Archaea to form the eukaryotes. Later, divides, copies of the same genome DOI 10.1002/bies.201500149 endosymbiosis of cyanobacteria with are found in each daughter cell. Ulti- eukaryotes led to green algae and mately, after populations of organisms plants. Many other gene transfers diverge or “speciate,” evolution along 1) Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA, together with multiple other endosym- each branch will leave genomic signals USA bioses have been inferred. Microbiolo- of that branching event in daughter 2) Department of Genetics, Evolution and gists began to argue that the “tree” of lineages. Sex and recombination can Environment, University College London, London, UK life was more like a web or network than obscure this picture, but in both 3) Department of Biological Sciences and Eck a tree [3–5]. Bacteria and Archaea sex (in the Institute for Global Health, University of Notre Today, whole genome sequencing is eukaryote-like sense of homologous Dame, Notre Dame, IN, USA providing unprecedented phylogenetic gene exchange) is mostly a transaction 4) Department of Biology and School of Informatics and Computing, Indiana University, information about whole groups of between closely related individuals, Bloomington, IN, USA eukaryotes [6–14]. Here we review mostly within the same populations or genomic evidence suggesting that retic- “species” [19–22]. Eukaryotes are simi- *Corresponding author: James Mallet ulate evolution may have considerable lar [23, 24]. Lateral transfer involving E-mail: [email protected] impact in multicellular eukaryotes as non-homologous exchange, on the

140 www.bioessays-journal.com Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited...... Insights & Perspectives J. Mallet et al.

other hand, will lead to more wide- derided as “the tree of one per pathologies to those found in prokar- ranging phylogenetic discordance. In cent” [32]. Around 80% of eukaryotic yotes. This raises doubt about the prokaryotes, both sex and lateral trans- proteins are actually more closely eukaryotic Tree of Life as a whole. Apart hn again Think fer involve relatively few genes at a time related to homologs in the Bacteria from phylogenetic estimation error and or even if more extensive, usually much than in the Archaea; the Universal homoplasy, there are three main causes less than 50% of the genome. Neverthe- Tree’scloserarchaeal-eukaryoteaffin- of phylogenetic incongruence: lateral less, multiple exchanges may take ityisreflectedinonlyabout15%of gene transfer, incomplete lineage sort- place, and very large fractions of eukaryote proteins [28, 32], including ing, and introgression. the genome might eventually be those used by Ciccarelli et al. [30]. exchanged with other lineages or spe- Because of concerns such as these, the cies over long periods. If so, it is possible existence of species and of the Univer- Lateral transfer that the signals of the organismic sal Tree in microbes has been dis- genealogy (the original “tree of cells”) missed as a “myth” in the prokaryote In Eukaryotes, lateral or horizontal in the genome will be obliterated by literature [33]. Whether species or the gene transfer is widespread, but is multiple phylogenetic signals from sex Universal Tree exist in prokaryotes has usually thought to be rare compared and lateral transfer. become almost a philosophical rather to that in prokaryotes [8, 36, 37]. It Before assessing new genomic evi- than a biological issue [29], but it does seems to be associated mainly with dence for phylogenetic discordance in seem clear that most of the original single-celled eukaryotes (the “pro- multicellular eukaryotes, it is worth Universal Tree, whether identifiable or tists”), especially those that engulf reviewing the controversy raging about not, is located on the far side of what their food, or in multicellular organ- the microbial “Tree of Life” over the last Woese originally intended by the isms with parasites in close cellular few decades. Carl Woese [25] argued Darwinian Threshold. contact with their hosts. Eukaryotes that in spite of considerable lateral clearly seem to have acquired impor- transfer, there is “a genealogy-defining tant genes via lateral transfer from both core of genes whose common history What causes phylogenetic mitochondrial and chloroplast endo- dates back to the root of the universal symbionts, but transfers also originate tree.” Woese suggested that the acqui- incongruence in from other endosymbionts, parasites, sition of sufficient co-adaptation among eukaryotes? and close associates [35]. Lateral pro- these key genes caused life to reach a cesses in eukaryotes, in contrast to “Darwinian threshold,” which permit- Findings of promiscuous gene exchange other possible causes of reticulation, ted divergence into separate species and among prokaryotes have usually may transfer genes between distantly allowed us to trace the organismal been contrasted with supposedly well- related species, but typically involve history, even while lateral transfer behaved trees in eukaryotes [33, 34]. relatively few genes at a time, as in obscures the universal tree for many Eukaryote genomes originated when an prokaryotes. Lateral transfer is com- other genes. According to Woese, before archaeal cell acquired many bacterial mon in some multicellular groups [36], the Darwinian threshold was reached, genes, in part but certainly not only such as bdelloid rotifers, which, inter- divergence and speciation could not associated with the bacterial endosym- estingly, lack meiotic sex [38, 39]. take place, and no tree of genes would biotic origins of mitochondria and Horizontal gene transfer in the mito- allow us to trace the organismal history. chloroplasts [35]. Eukaryotes also chondrial genomes of plants and yeasts It quickly became apparent that invented meiosis, which allows recom- is also widespread [40]. However, lateral transfer does indeed swamp the bination of whole genomes. In multicel- horizontal transfer is probably not an signal of the Universal Tree in micro- lular eukaryotes, reproduction itself overriding factor in the evolution of the bial genomes: in fact no other genes often involves meiosis. This innovation nuclear protein coding genes of most support Woese’s original 16S RNA effectively destroys the tree-like signal multicellular eukaryotes, unlike those tree [26]. Many microbiologists now in an organismal (“tree of cells”) of prokaryotes. deny a tree-like phylogeny of micro- phylogeny. In every meiosis recombi- In contrast to the genes, eukaryotic bial evolution; instead the phylogeny nant haploid genomes from two suc- genomes often consist largely of non- of life looks more like a web or a ring cessful, independent cells are thrown coding DNA, and 30–60% of this [3, 27–29]. By excluding all genes that together to form diploid zygotes, before consists of recognizable mobile ele- disagree with the Universal Tree, one the sum of the genetic material is ments [41, 42]. Intergenic and intronic can select 20–30 largely informational haphazardly and approximately equally DNA is thought to originate largely via genesthatmoreorlessrescuethe recombined into haploid daughter cells. active or inactivated mobile genetic ribosomal RNA Tree [29–31]. But this A “tree of cells” justification for the elements [43–45], most of which are almost seems like cheating, and is eukaryote Tree of Life is no longer thought to enter lineages via lateral itself obtained only by pruning out a possible. transfer [46]. Mobile elements are par- number of clear cases of lateral While tree-like patterns are readily ticularly likely to be important in the transfer in even these genes. As this discernible in eukaryote phylogenies, evolution and spread of regulatory anyway only applies to a tiny fraction we here highlight recent evidence elements. Nonetheless, the introduction of the genome, these recent incarna- suggesting that a number of regions of of new mobile elements via lateral tionsoftheUniversalTreehavebeen the eukaryotic tree show similar transfer is rare, and the lifespan of

Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc. 141 J. Mallet et al. Insights & Perspectives.....

active proliferation via transposition is trees has been shown to converge on an related groups of species, unlike lateral cut short by relatively rapid loss, incorrect but highly signiﬁcant solu- gene transfer. This is a major difference inactivation and sequence degradation tion [52, 53]. In spite of this “tyranny of between reticulate evolution in prokar- in the host genome [46]. the majority” in phylogenetic analysis, a yotes and eukaryotes: while lateral gene coalescent-based analysis should none- transfer weaves lineages together across theless be able to retrieve the true disparate parts of a tree, introgression Incomplete lineage sorting bifurcation signal in spite of the con- merely results in tangled knots on a local fused gene tree signal [54, 55]. scale. Nonetheless, introgression has The two main causes of gene tree – potentially important effects throughout species tree discordance, at least for the tree of life by obscuring relationships

Think again protein-coding genes in closely related Introgression and reticulated amonglineages that diversified rapidly at groups of eukaryotes, are incomplete evolution any time, not just in those that did so lineage sorting and introgression. In- recently. complete lineage sorting occurs when The third source of phylogenetic incon- Introgression was well known be- polymorphisms persist between specia- gruence, introgression, occurs when fore the advent of genetic markers or tion events, so that the actual (true) hybrids backcross and transfer genetic genomics, and was long believed an genealogical relationship of a gene or material between species. Hybridization important catalyst for adaptive evolu- genome region differs from the true may occur without strongly affecting tionary change in plants [16]. Introgres- species branching pattern. As an exam- the genomes of recipient populations if sion was thus familiar by the 1960s, but ple of incomplete lineage sorting, strongly resisted by selection, but ideas of “coadapted gene complexes,” around 15% of human genes are more genomic admixture results if the intro- and “the unity of the genotype” associ- closely related to homologs in gorillas gressed alleles are established. ated with the biological species concept than to those in our true sister lineage, Hybridization between related eu- led to a belief that hybridization had the chimpanzees, while another 15% of karyote species does occur reasonably little importance in animals, at least. genes group gorilla and chimpanzee. frequently in nature; it is known to affect When hybridization did occur, it was This is expected from what we know around 25% of the species of flowering often assumed to be unnatural and about the ancestral effective population plants and about 10% of animals [56–58]. was attributed to environmental sizes of these species and the short time The fraction of hybrids in natural pop- changes wrought by humans [68]. between human-gorilla and human- ulations, nevertheless, is usually low: Because hybrids are mostly unfit, it chimpanzee speciation events [47, 48]. natural interspecific hybridization rates was assumed that introgression among In some cases, incomplete lineage in animals are typically 0.1% or less per animal species very rarely had any long- sorting occurs as a result of balancing generation in any species [57, 59]. Per term evolutionary impact [15]. selection maintaining polymorphisms: generation hybridization rates can be With the potential for introgression, when speciation occurs, both daughter much higher in some populations of not only will individual gene trees tell species may maintain the same “trans- plants and animals, where it reaches different stories, but the actual organis- specific” polymorphisms, even though several per cent, for example in some mal branching pattern between species with recombination, the signal of an- oaks (Quercus), Darwin’s finches, and will be reticulate, rather than strictly cestral origin may erode over time [49]. some cases in Heliconius butter- bifurcating. The true phylogeny may be Good examples of shared polymor- flies [60–63]; but these are probably approximately tree-like if introgression phisms between humans and apes are exceptional. While some hybrids are is rare and affects only a very small MHC [50] and ABO blood group loci [51], sterile, a substantial fraction of such fraction of the genome, but will not be among other genes. In the species hybrids are at least partly fertile, leading tree-like if introgression is common. complex including the major mosquito to observed cases of backcrossing and However, the importance of introgres- vector, Anopheles gambiae, a very large introgression. It is important to realize sion is only now becoming apparent chromosomal inversion, 2La (22 Mb in that hybridization and introgression may with rapid genome sequencing. length, 8.5% of the total genome size) is occur among non-sister species as well as maintained as a balanced polymor- between sister species, especially during phism that has persisted across several rapid adaptive radiations. Gene transfer is important speciation events [18]. Closely related species hybridize in eukaryote genomes Unlike lateral transfer and introgres- more readily than more distant spe- sion, however, discordance created by cies [64]. The decline of natural hybrid- The extent of introgression incomplete lineage sorting does not ization rates with genetic distance, while across the eukaryote tree imply phylogenetic reticulation at the noisy,maybeveryroughlyapproximated level of species. It merely muddles the as exponential [59], mirroring the noisy As we have seen, meiotic fertility has an genomic signal of what might be a truly decline of compatibility in meta-analyses increasing tendency to fail with genetic bifurcating phylogeny. In some trees of transformation experiments in prokar- distance, but failure is often not com- with four or more taxa and rapid yotes and laboratory crosses in animals plete in the closest hybrids. For this successive speciation events (the and plants [19–23, 64–67]. Thus, intro- reason, introgression, which requires “anomaly zone” of phylogenetics), gression tends to generate phylogenetic some fertile hybrid offspring, is most the species tree estimated from the gene discordance mainly among closely likely to occur among closely related

142 Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc...... Insights & Perspectives J. Mallet et al.

species. Hybridization between sister species will not usually affect the species tree topology, but will make hn again Think the apparent divergence time between the species appear more recent [69]. However, if two widely distributed species interact in populations where they overlap, it may be possible that individual populations become on average more closely related locally to a sister species than to more distant conspecific populations. In contrast, hybridization and introgression among non-sister lineages can readily distort Figure 1. A: “Whole genome” versus “species” tree topologies of the Anopheles gambiae the species tree topology. If introgres- complex in Africa. B: The tree based on the X chromosome only, showing introgression sion between non-sister lineages is events and estimated node divergence times. The average phylogeny of the whole genome widespread across the genome, it may is distorted by autosomal introgression between A. gambiae þ coluzzii and A. arabiensis, but be very hard to retrieve the true this was prevented on the X chromosome by X-linked hybrid incompatibilities and multiple bifurcation history of the species. This overlapping inversions that prevent recombination. Modified and reprinted from [18] with is because a unitary history of the permission from AAAS. genome may not exist; if inferred from multiple loci or whole genomes, this species tree may be meaningless or from A. gambiae to A. arabiensis [18]. example, per generation hybridization misleading. Here we discuss several Introgression is on-going, and is an rates among closely related species of recent examples from multicellular excellent explanation for the phyloge- Darwin’s finches can be as much as 6%, eukaryotes where this may have oc- netic discordance, because wild hybrids with high fertility of hybrids. The Dar- curred. Interestingly, most of these and backcrosses between the latter two win’s finches began to diversify on the examples come from rapid species species are 0.22% of the individuals Cocos and Galapagos Islands less than 1 radiations; these are exactly the cases captured in sympatry [71, 72]. In decid- million years ago, and there is strong in which closely related but non-sister ing between conflicting topologies, the genomic evidence for past and continu- species may be hybridizing. species tree was inferred from regions of ing introgression across almost the entire The group of eight African mosquito the genome with the deepest coales- group [74]. Other vertebrate groups such species known as the A. gambiae cence times between species [18]. If this as African lake cichlids, Xiphophorus complex radiated within the last 2 information had not been available, or if fishes, horses, and even hominins show million years [18]. Species distributions introgression had been even more similar phylogenetic discordance in- overlap extensively, and in areas of complex, it would have been hard to ferred to be due to introgression [75–78]. sympatry hybrids have been recorded at infer the species tree at all. Much deeper evidence of reticulate rates of 0.02–0.75% [70, 71]. Despite In Heliconius butterflies, the “mel- evolutionary patterns also exists. For F1 hybrid male sterility in most cases, pomene-silvaniform” clade consists of example, there is considerable phyloge- introgression is plausible through the around 15 species. Most of these are netic discordance at the base of the backcrossing of vigorous and fertile F1 “good” species that co-occur over large Neoaves, or modern birds [79–81]. In fact hybrid females. When genomes of sympatric regions, and are somewhat none of the thousands of individual gene multiple members of the A. gambiae interfertile with other members of the trees support the various conflicting complex were sequenced and com- clade. However, rare hybrids and back- estimates of the species tree [79, 81]. pared, the inferred species tree was crosses are known from the wild and in Trees built from indels and stable mobile evident in only 2% of the genome, captivity across this whole group, element insertions (which are less prone mainly on the X chromosome, whereas suggesting the possibility that a slow to homoplasy than nucleotide or amino the majority tree in the rest of the trickle of introgression is constantly acid substitutions) show similar conflict, genome yields a completely contradic- occurring among the largely sympatric suggesting that thegene tree discordance tory tree [18]. While some of these species in the group [59]. This sugges- is real, rather than due to phylogenetic differences are due to incomplete line- tion has now been confirmed: because error [79]. The authors of these papers age sorting, much of this discordance is of introgression, a local population of argued that the tangle at the base of this due to introgression between two non- H. melpomene can be more closely ancient radiation was due to incomplete sister species (Fig. 1). This is particularly related to the locally overlapping popu- lineage sorting, but did not address the clear for the 2La inversion mentioned lation of its sister H. cydno than it is possibility of introgression. Yet introgres- above, which is inferred to have been to conspecifics at over 40% of the sion seems a likely additional cause: polymorphic in the ancestor of the genome [17, 73]. around 9% of today’s bird species are complex, but is affected by three losses Rapidradiations suchasthesetendto known to hybridize in the wild [56], and of 2Lþ and one of 2La, as well as one produce many closely related species birds retain some hybrid compatibility fairly recent (1 Mya) introgression of 2La that may be partially interfertile. For with congeners for 10 My after

Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc. 143 J. Mallet et al. Insights & Perspectives.....

speciation [65]. After the demise of the parents [17]. That case remains to be their genealogical history. The true dinosaurs, the early Neoaves had few fully worked out, but similar cases have phylogeny will trace the disparate competitors, and itisnotunlikelythatthe been put forward for cichlid fish, histories of every gene, and cannot first species in today’s lineages were able monkeyflowers, and other hybridizing readily be represented on a page, to hybridize with one another during adaptive radiations [86, 87]. In one case, certainly not as a single tree. Yet their global ecological diversification, the beginnings of the process have been we propose that there may still be a much as the Darwin’s finches do today observed in “real time”: a new hybrid true bifurcating tree of species out on the Galapagos Islands [74]. An finch species that breeds strictly endo- there (Fig. 3), in spite of the back- explanation for the strong signals of gamously has now been followed on a ground chaos of gene trees. Only if discordance at the base of the Neo- Galapagos island for seven generations species fuse either wholly or in some

Think again aves [79, 81] may therefore lie partly in since its formation via initial hybridiza- geographic region to become a single gene flow among the lineages after they tion events in the early 1980s [88]. cluster (e.g. in sticklebacks [90] or in diversified. Given that other major hybrid speciation), does the species groups, such as the placental mam- phylogeny itself become reticulate mals [13], or the animals as a whole [82], under this view. appear also to have evolved in rapid Introgression challenges Possible alternatives to the species radiations, it seems likely that our notions of species and tree is some consensus of gene trees, or persistent problems with estimation of phylogeny perhaps the tree based on the “demo- trees for the deepest branches of these cratic majority” of the genome [91]. radiations is due to historical introgres- The meaning of species and Obtaining the maximum likelihood or sion as well as incomplete lineage sorting speciation most probable species tree from a series during their initial diversification. of genes is in fact the aim of many We are thus confronted by extraordi- phylogenetic and phylogenomic stud- nary levels of introgression found in the ies, at least among eukaryote system- Is introgression adaptive? genomes of rapidly radiating species atists [92, 93]. This program assumes (such as Anopheles, Heliconius, and that the true species tree is more likely Phylogenetic or genealogical studies of Darwin’s finches). Yet these taxa are to emerge via analysis of larger fractions the extent of introgression across the currently readily identifiable to species of the genome. Under the viewpoint genome do not, however, reveal using morphology or genetics: none of proposed here, this is not necessarily whether the process is largely neutral us doubt that the species is a useful true if there is abundant introgression or whether it is aided by a selective rank, at least in multicellular eukar- (Fig. 3). For example, as shown above, advantage on the new genomic back- yotes. We recognize these taxa as the single most common tree inferred ground. The relative importance of species not because of reproductive from whole genomes of the Anopheles selection in introgression across the isolation per se, nor because they gambiae complex in Africa gives an genome is still not known, and is an represent phylogenetic branching incorrect rendering of the group’s area of active research [83], but many events, but because of the simpler history [18] (Fig. 1). introgression events are now known to observation that hybrids and intermedi- Historical introgression events in have involved adaptation. A number of ates between the clusters we call taxa such as Anopheles have been transfers of mimicry-determining loci species [89] are rare. While most of inferred to affect the majority of the have been documented in Heliconius the introgression that has resulted in genome, even though natural hybrids butterflies (Fig. 2A and B), and in reticulate relationships occurred in the are relatively rare among the contem- Anopheles the many cases of insecticide past – and may or may not be ongoing – porary species (see above for rates of resistance alleles crossing species these results suggest that species are hybridization). Nonetheless, hybrid- boundaries (see below) and the exis- like the Ship of Theseus in philosophy, ization can introduce variation at rates tence of balancing selection at the 2La which can progressively but almost much higher than mutation, so that inversion make it rather hard to believe completely be rebuilt with new wood, significant levels of genomic replace- that selection is only rarely involved in and yet remain the same ship. We do not ment may accrue over long periods, introgression. yet know how common these effects are even at the low hybridization rates Adaptive introgression may also among genomes of other eukaryotes, known in Anopheles today. Similar introduce adaptive combinations that but the recent discoveries in mammals, results also apply in some Heliconius lead to new species, or hybrid specia- birds, fish, insects, plants, and fungi species. If we wish the species tree to tion [84, 85]. Plant examples have long suggest that they may be widespread be determined by the democratic been known [16, 85], but animal exam- throughout the eukaryotic Tree of Life. opinion of the genes, we are therefore ples are no longer rare. For example, the forced to accept a peculiar species Heliconius pardalinus-like ancestor of definition that perhaps applies only to H. elevatus seems to have recently The “true phylogeny” versus terminal taxa, rather than the original acquired the majority of its defensive the species tree bifurcating ancestors, because the color pattern mimicry from H. melpom- branchesofthetreechangetheir ene (Fig. 2C), subsequently proving able In introgressing species, different gene species identity whenever accumula- to coexist in sympatry with both trees vary in the story they tell about tion of introgressed regions flips the

144 Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc...... Insights & Perspectives J. Mallet et al. hn again Think

Figure 2. Phylogenetic discordance B/D mimicry region of Heliconius genomes. A: FST plot Are species incompatible? shows divergent optix regulatory region determining mimicry differences between geographic races within H. melpomene. Mimicry has been shown to have very strong adaptive value in Another conclusion that arises from Heliconius. B: The same region shows a strong excess of ABBA phylogenetic sites over these findings is that large fractions of BABA sites, implicating introgression between H. melpomene and H. timareta. C: Further- different species’ genomes may in fact more, the non-sister species H. elevatus shows a phylogenetic topology indicating be compatible. The genomic distribu- introgression of the rayed mimicry pattern from the melpomene-timareta clade in the same tion of “intrinsic” incompatibilities genomic region. Modified and reprinted with permission from [17]. (such as “Dobzhansky-Muller incompatibilities” [94, 95]) is poorly known except in a few species [96]. In Saccha- democratic majority of the genes to theideathatthespeciestreeisthe romyces yeasts, it is possible to replace another topology. It is perhaps defen- bifurcation history (Fig. 3). This we whole chromosomes with little effect on sible to argue that the “democratic would argue is closer to what we mean viability, while in Drosophila many opinion” tree is more predictive of the bythespeciationhistory,inspiteof hybrid sterility loci seem scattered very origins of the genes, though it is the difficulty of its discovery, and widely across the genome [97–101]. It is marred by potential inferences of acknowledging a lowered expectation possible that the situation in Drosophila ancestral species (pale green) that of its predictiveness for the histories of is unusual, perhaps a result of “faster never existed (Fig. 3). We instead favor its component genes. male” sexual selection that leads to

Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc. 145 J. Mallet et al. Insights & Perspectives.....

individuals seeking sexual partners, albeit with different parameter values. If outcrossing within and between species is regulated by the same cost/ beneﬁt equation, a sexual selection process similar to reinforcement should apply to interactions within as well as between species.

Think again Practical implications of introgression The prevalence of laterally transferred antibiotic resistance genes among bacterial species is a well-known problem for human health [108, 109]. Similar problems might therefore be expected to result from introgression or lateral transfer among related eukaryotic pest and disease species. The African malaria-carrying mosquitoes provide some Figure 3. A simple case where introgression can distort the history of species and worrying examples. For example, rates speciation. By “the true phylogeny,” in this paper, we mean the totality of true histories of of hybridization between Anopheles every part of the genome. This is not readily depicted: our simplified cartoon of the true phylogeny network above indicates abundant introgression between species 1 and 2 after gambiae and A. arabiensis are only their bifurcation, but little between sister species 2 and 3. It does not, however, show which 0.22% per generation [71]. However, gene travels in which direction and when, all of which is surely important information about because this introduces foreign alleles the “true phylogeny” as well. If introgression is extensive, the whole genome tree (bottom left) at a rate far higher than mutation, there may indicate an incorrect bifurcation history, as well as ancestral species that never existed are persistent concerns that insecticide (such as the apparent ancestor of 1 and 2 in the diagram). The true bifurcation history of resistance evolution in one species may species is shown bottom right. lead to the rapid spread of that resistance to others via introgression [72]. Multiple cases of introgression of alleles genome-wide effects on male hybrid adequate to assess the potential for encoding both organophosphate and sterility [102, 103]. Even though incom- introgression between species: we will pyrethroid insecticide resistance are patibility loci have been mapped in also need to know the number and certainly known between the sister crosses between A. gambiae and A. selective effects of these variants. species A. gambiae and A. coluz- arabiensis [104], genomic evidence for As far as is known, classical lateral zii [110–113]; these two are known to very widespread homologous replace- transfer from distantly related species is hybridize and backcross much more ment between species in the autosomes not a major recent source of phyloge- frequently [63] than do A. gambiae with of Anopheles and Heliconius [18, 73] netic incongruence in multicellular A. arabiensis. Similarly, large sibling suggests either that incompatibilities eukaryotes, and most of the phyloge- species complexes of the black fly genus were not very common in those netic reticulation we observe is due to Simulium transmit river blindness in genomes, or that some introgressed homologous exchange via hybridiza- Africa and tropical America, and may alleles are advantageous enough to tion. The selective advantages of sex also exchange genes. Among sympatric overcome initial incompatibility. Al- within species remain contentious, but species of the African S. damnosum though autosomal genes introgress sex surely optimizes some balance complex, hybridization rates may reach readily in both groups, the preponder- between benefits and costs of recombi- 0.1% per generation. Introgression is ance of “species tree” genealogies in the nation [105, 106]. Typically, hybridizing thought likely to explain the rapid sex chromosome in the Anopheles with another species is viewed as “the spread of insecticide resistance among gambiae complex [18] is likely due to grossest blunder in sexual preference,” multiple Simulium species in Africa [114]. multiple overlapping inversions that and mate choice (reinforcement) is The same problem occurs even in differ between A. gambiae þ coluzzii expected to evolve to limit hybridization vertebrate pests: a genomic region and A. arabiensis. These inversions among sympatric species [107]. How- containing a rodenticide resistance suppress recombination and so inhibit ever, given that hybridization does still allele spread via introgression between introgression of small chromosomal occur, and sometimes leads to benefi- two partially interfertile mouse species fragments on this chromosome. If cial effects, we should now perhaps in Western Europe [115]. adaptive alleles are widely available to broaden our view of sex across the Recent advances in genetics and introgress, determining the number and species boundary, where the same genetic engineering are revolutionizing effect of incompatibilities will not be cost/benefit function is confronted by pest control, allowing for “designer

146 Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc...... Insights & Perspectives J. Mallet et al.

organisms” in agriculture and human groups of rapidly radiating species may funding from the Broad Insititute of health. Several major transgenic crops, exchange adaptive as well as non- Harvard and MIT, and NIH grant R01 especially those expressing herbicide or adaptive genomic regions, as in Heli- AI76584. hn again Think insect resistance, have been released in conius, Anopheles, cichlids, Xiphopho- The authors have declared no conflicts many countries. At the same time, new rus, Darwin’s finches, horses, and of interest. molecular marker and genomic analy- hominins. In fact, because hybridiza- ses let us gather evidence on gene flow tion between sister species does not between crops and wild relatives for the always affect the species tree – and first time. The results are clear: intro- because introgression between sister References gression does occur, and weedy rela- species is more likely – it may be that 1. Woese CR, Kandler O, Wheelis ML. 1990. tives are acquiring novel genetic estimates of introgression rates from Towards a natural system of organisms: variation from crops, including trans- species tree topologies alone vastly proposal for the domains Archaea, Bacteria, genes that are liable to make these underestimate the amount of gene flow and Eucarya. Proc Natl Acad Sci USA 87: 4576–9. weeds more noxious [116]. occurring in nature. For many systems 2. Woese CR, Fox GE. 1977. Phylogenetic The use of transgenic organisms is we may think we are able to infer a structure of the prokaryotic domain: the more advanced in agriculture than in species tree signal, but we must recog- primary kingdoms. Proc Natl Acad Sci USA human health. However, a variety of nize that this signal may only be 74: 5088–90. 3. Doolittle WF. 1999. Phylogenetic classifi- genetic control measures of vectors represented by a small fraction of genes. cation and the universal tree. Science 284: have been suggested and in some cases As well as causing problems for 2124–9. are being used to engineer disease phylogenetics, abundant introgression 4. Hilario E, Gogarten JP. 1993. Horizon- tal transfer of ATPase genes – the tree of vector populations [117]. For example, and incomplete lineage sorting might life becomes a net of life. BioSyst 31: infection of Aedes mosquitoes by Wol- greatly weaken inferences in compara- 111–9. bachia causes refractoriness to dengue tive analysis. When we map character 5. Martin W. 1999. Mosaic bacterial chromo- virus proliferation [118], while Wolba- traits onto the tree of a rapidly radiating somes: a challenge en route to a tree of genomes. BioEssays 21: 99–104. chia-infected Anopheles mosquitoes group, we should be cautious. For 6. Clark AG, Eisen MB, Smith DR, Bergman show reduced Plasmodium infec- instance, the raptorial habit is thought CM,etal. 2007. Evolution of genes and tion [119]. In addition, there is the to be ancestral to the entire core land- genomes on the Drosophila phylogeny. Nature 450: 203–18. possibility of manipulating the genetics birds, but today it is present in several 7. Lindblad-Toh K, Garber M, Zuk O, Lin MF, of mosquito innate immunity in order to monophyletic groups, each more et al. 2011. A high-resolution map of human reduce their efficiency as a vector [120]. closely related to birds that have evolutionary constraint using 29 mammals. Nature 478: 476–82. Of these, probably the most successfully apparently lost the habit [79–81]. Alter- 8. Dujon B, Sherman D, Fischer G, Durrens used cases so far are a number of natively, core landbirds may have been P,etal. 2004. Genome evolution in yeasts. releases of Wolbachia-infected Aedes ancestrally non-raptorial, and a number Nature 430: 35–44. aegypti to control dengue ([118] www. of raptorial traits could have could have 9. Maia VH, Gitzendanner MA, Soltis PS, Wong GK-S,etal. 2014. Angiosperm eliminatedengue.com). As with trans- been shared at the base of these lineages phylogeny based on 18S/26S rDNA se- genic plants, because the transmission by introgression among the early spe- quence data: constructing a large data set of genetic traits requires mating, these cies of each lineage. This is not using next-generation sequence data. Int J Plant Sci 175: 613–50. genetic traits may “leak” to related dissimilar to what we observe in mim- 10. Lee EK, Cibrian-Jaramillo A, Kolokotronis species via introgression. This may not icry patterns in Heliconius or in beak S-O, Katari MS,etal. 2011. A functional have negative impacts, especially in morphology of Darwin’s finches, among pylogenomic view of the seed plants. PLoS comparison to the potentially positive species of radiations that we see hy- Genet 7: e1002411. 11. Marcussen T, Sandve SR, Heier L, Span- benefits of the engineered trait on the bridizing today [17, 74]. Similarly, nagl M,etal. 2014. Ancient hybridizations target species. However, given genomic inferences from phylogeography – such among the ancestral genomes of bread evidence for introgression of many other as geographic origins of rapidly radiat- wheat. Science 345: 1250092. 12. Brandvain Y, Kenney AM, Flagel L, Coop traits, its importance should not be ing groups inferred from phylogenetic G,etal. 2014. Speciation and introgression underemphasized when seeking regula- methods – should be affected as well. between Mimulus nasutus and Mimulus tory approval for release of genetically The origins of traits, and the genes that guttatus. PLoS Genet 10: e1004410. 13. McCormack JE, Faircloth BC, Crawford modified organisms (recognizing that determine them can have very different NG, Gowaty PA,etal. 2012. Ultraconserved Wolbachia infection does not techni- histories from that of the species tree. elements are novel phylogenomic markers cally qualify as a genetic modification to that resolve placental mammal phylogeny most regulatory bodies). when combined with species-tree analysis. Genome Res 22: 746–54. Acknowledgments 14. Yang Y, Moore MJ, Brockington SF, Soltis We thank Kanchon Dasmahapatra, Bill DE,etal. 2015. Dissecting molecular Conclusions and outlook Hanage, Robin Hopkins, James McIner- evolution in the highly diverse plant clade Caryophyllales using transcriptome se- ney, Jesse Shapiro, Leonie Moyle, Luay quencing. Mol Biol Evol 32: 2001–14. Our main conclusion is that many more Nakhleh, Tim Sackton, and Ziheng Yang 15. Mayr E. 1963. Animal Species and Evolu- species are likely exchanging genes for discussions. This paper benefited tion. Cambridge, MA: Harvard University than has been appreciated. It is not from the comments of three anonymous Press. 16. Anderson E, Stebbins GL. 1954. Hybrid- only sister species that hybridize and reviewers. The authors were supported ization as an evolutionary stimulus. Evolution undergo genomic introgression: whole by BBSRC, Harvard University, SPARC 8: 378–88.

Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc. 147 J. Mallet et al. Insights & Perspectives.....

17. Heliconius Genome Consortium. 2012. But- of both vertebrate and invertebrate 58. Arnold ML, Sapir Y, Martin NH. 2008. terfly genome reveals promiscuous ex- genomes. Genome Biol 16: 50. Genetic exchange and the origin of adapta- change of mimicry adaptations among 38. Gladyshev EA, Meselson M, Arkhipova IR. tions: prokaryotes to primates. Philos Trans species. Nature 487: 94–8. 2008. Massive horizontal gene transfer in R Soc Lond B Biol Sci 363: 2813–20. 18. Fontaine MC, Pease JB, Steele A, Water- bdelloid rotifers. Science 320: 1210–3. 59. Mallet J, Beltran M, Neukirchen W, house RM,etal. 2015. Extensive introgres- 39. Boschetti C, Carr A, Crisp A, Eyres I,etal. Linares M. 2007. Natural hybridization in sion in a malaria vector species complex 2012. Biochemical diversification through heliconiine butterflies: the species boundary revealed by phylogenomics. Science 347: foreign gene expression in bdelloid rotifers. as a continuum. BMC Evol Biol 7: 28. 1258524. PLoS Genet 8: e1003035. 60. Bacilieri R, Ducousso A, Petit RJ, Kremer 19. Zawadzki P, Roberts MS, Cohan FM. 40. Wu B, Buljic A, Hao W. 2015. Extensive A. 1996. Mating system and asymmetric 1995. The log-linear relationship between horizontal transfer and homologous recom- hybridization in a mixed stand of European sexual isolation and sequence divergence in bination generate highly chimeric mitochon- oaks. Evolution 50: 900–8. Bacillus is robust. Genetics 140: 917–32. drial genomes in yeast. Mol Biol Evol 32: 61. Grant PR. 1993. Hybridization of Darwin’s Think again 20. Majewski J, Zawadzki P, Pickerill P, 2559–70. finches on Isla Daphne Major, Galapagos. Cohan FM,etal. 2000. Barriers to genetic 41. Sela N, Kim E, Ast G. 2010. The role of Philos Trans R Soc Lond B Biol Sci 340: exchange between bacterial species: Strep- transposable elements in the evolution of 127–39. tococcus pneumoniae transformation. non-mammalian vertebrates and inverte- 62. Mallet J, McMillan WO, Jiggins CD. 1998. J Bacteriol 182: 1016–23. brates. Genome Biol 11: R59. Estimating the mating behavior of a pair of 21. Eppley JM, Tyson GW, Getz WM, Banfield 42. Kidwell MG. 2002. Transposable elements hybridizing Heliconius species in the wild. JF. 2007. Genetic exchange across a and the evolution of genome size in Evolution 52: 503–10. species boundary in the archaeal genus eukaryotes. Genetica 115: 49–63. 63. Lee Y, Marsden CD, Norris LC, Collier TC, Ferroplasma. Genetics 177: 407–16. 43. Lynch M. 2007. The Origins of Genome et al. 2013. Spatiotemporal dynamics of 22. Naor A, Lapierre P, Mevarech M, Papke Architecture. Sunderland, MA: Sinauer gene flow and hybrid fitness between the M RT,etal. 2012. Low species barriers in Associates. and S forms of the malaria mosquito, halophilic Archaea and the formation of 44. Lynch M. 2006. The origins of eukaryotic Anopheles gambiae. Proc Natl Acad Sci recombinant hybrids. Curr Biol 22: 1444–8. gene structure. Mol Biol Evol 23: 450–68. USA 110: 19854–9. 23. Gourbiere S, Mallet J. 2010. Are species 45. Rogozin IB, Carmel L, Csuros M, Koonin 64. Coyne JA, Orr HA. 1997. “Patterns of real? The shape of the species boundary EV. 2012. Origin and evolution of spliceo- speciation in Drosophila” revisited. Evolution with exponential failure, reinforcement, and somal introns. Biol Direct 7: 11. 51: 295–303. the “missing snowball”. Evolution 64: 46. Schaack S, Gilbert C, Feschotte C. 2010. 65. Price TD, Bouvier MM. 2002. The evolution 1–24. Promiscuous DNA: horizontal transfer of of F1 postzygotic incompatibilities in birds. 24. Morales L, Dujon B. 2012. Evolutionary role transposable elements and why it matters Evolution 56: 2083–9. of interspecies hybridization and genetic for eukaryotic evolution. Trends Ecol Evol 66. Moyle LC, Nakazato T. 2010. Hybrid exchanges in yeasts. Microbiol Mol Biol Rev 25: 537–46. incompatibility “snowballs” between Sola- 76: 721–39. 47. Burgess R, Yang Z. 2008. Estimation of num species. Science 329: 1521–3. 25. Woese CR. 2002. On the evolution of cells. hominoid ancestral population sizes under 67. Moyle LC, Olson MS, Tiffin P, Baum D. Proc Natl Acad Sci USA 99: 8742–7. Bayesian coalescent models incorporating 2004. Patterns of reproductive isolation in 26. Creevey CJ, Doerks T, Fitzpatrick DA, mutation rate variation and sequencing three angiosperm genera. Evolution 58: Raes J,etal. 2011. Universally distributed errors. Mol Biol Evol 25: 1979–94. 1195–208. single-copy genes indicate a constant rate 48. Dutheil JY, Munch K, Nam K, Mailund T, 68. Anderson E. 1948. Hybridization of the of horizontal transfer. PLoS ONE 6: e22099. et al. 2015. Strong selective sweeps on the X habitat. Evolution 2: 1–9. 27. Rivera MC, Lake JA. 2004. The ring of life chromosome in the human-chimpanzee 69. LeacheAD , Harris RB, Rannala B, Yang Z. provides evidence for a genome fusion ancestor explain its low divergence. PLoS 2013. The influence of gene flow on species origin of eukaryotes. Nature 431: 152–5. Genet 11: e1005451. tree estimation: a simulation study. Syst Biol 28. McInerney JO, O’Connell MJ, Pisani D. 49. Wiuf C, Zhao K, Innan H, Nordborg M. 63: 17–30. 2014. The hybrid nature of the Eukaryota 2004. The probability and chromosomal 70. Coluzzi M, Sabatini A, Petrarca V, Di Deco and a consilient view of life on Earth. Nat Rev extent of trans-specific polymorphism. Ge- MA. 1979. Chromosomal differentiation and Microbiol 12: 449–55. netics 168: 2363–72. adaptation to human environments in the 29. PuigboP , Wolf YI, Koonin EV. 2009. 50. Teixeira JC, de Filippo C, Weihmann A, Anopheles gambiae complex. Trans R Soc Search for a ‘Tree of Life’ in the thicket of Meneu JR,etal. 2015. Long-term balancing Trop Med Hyg 73: 483–97. the phylogenetic forest. J Biol 8: 1–17. selection in LAD1 maintains a missense 71. Weetman D, Steen K, Rippon EJ, Mawejje 30. Ciccarelli FD, Doerks T, von Mering C, trans-species polymorphism in humans, HD,etal. 2014. Contemporary gene flow Creevey CJ,etal. 2006. Toward automatic chimpanzees, and bonobos. Mol Biol Evol between wild An. gambiae s.s. and An. reconstruction of a highly resolved tree of 32: 1186–96. arabiensis. Parasites Vectors 7: 345. life. Science 311: 1283–7. 51. Saitou N, Yamamoto F. 1997. Evolution of 72. Mawejje HD, Wilding CS, Rippon EJ, 31. Lang JM, Darling AE, Eisen JA. 2013. primate ABO blood group genes and their Hughes A,etal. 2013. Insecticide resis- Phylogeny of bacterial and archaeal homologous genes. Mol Biol Evol 14: tance monitoring of field-collected Anophe- genomes using conserved genes: super- 399–411. les gambiae s.l. populations from Jinja, trees and supermatrices. PLoS ONE 8: 52. Degnan JH, Rosenberg NA. 2006. Discor- eastern Uganda, identifies high levels of e62510. dance of species trees with their most likely pyrethroid resistance. Med Vet Entomol 27: 32. Dagan T, Martin W. 2006. The tree of one gene trees. PLoS Genet 2: e68. 276–83. percent. Genome Biol 7: 118. 53. Yang Z. 2014. Molecular Evolution. A 73. Martin SH, Dasmahapatra KK, Nadeau 33. Lawrence JG, Retchless AC. 2010. The Statistical Approach. Oxford: Oxford Uni- NJ, Salazar C,etal. 2013. Genome-wide myth of bacterial species and speciation. versity Press. evidence for speciation with gene flow in Biol Philos 25: 569–88. 54. Yang Z, Rannala B. 2014. Unguided Heliconius butterflies. Genome Res 23: 34. Mindell DP. 2013. The Tree of Life: meta- species delimitation using DNA sequence 1817–28. phor, model, and heuristic device. Syst Biol data from multiple loci. Mol Biol Evol 31: 74. Lamichhaney S, Berglund J, Almen MS, 62: 479–89. 3125–35. Maqbool K,etal. 2015. Evolution of 35. Huang J. 2013. Horizontal gene transfer in 55. Degnan JH, Rosenberg NA. 2009. Gene Darwin’s finches and their beaks revealed eukaryotes: the weak-link model. BioEssays tree discordance, phylogenetic inference by genome sequencing. Nature 518: 371–5. 35: 868–75. and the multispecies coalescent. Trends 75. Brawand D, Wagner CE, Li YI, Malinsky M, 36. Keeling PJ, Palmer JD. 2008. Horizontal Ecol Evol 24: 332–40. et al. 2014. The genomic substrate for gene transfer in eukaryotic evolution. Nat 56. Grant PR, Grant BR. 1992. Hybridization of adaptive radiation in African cichlid fish. Rev Genet 9: 605–18. bird species. Science 256: 193–7. Nature 513: 375–81. 37. Crisp A, Boschetti C, Perry M, Tunnacliffe 57. Mallet J. 2005. Hybridization as an inva- 76. Jonsson H, Schubert M, Seguin-Orlando A,etal. 2015. Expression of multiple sion of the genome. Trends Ecol Evol 20: A, Ginolhac A,etal. 2014. Speciation with horizontally acquired genes is a hallmark 229–37. gene flow in equids despite extensive

148 Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc...... Insights & Perspectives J. Mallet et al.

chromosomal plasticity. Proc Natl Acad Sci 92. Liu L, Yu L, Kubatko L, Pearl DK,etal. 108. Davies J, Davies D. 2010. Origins and USA 111: 18655–60. 2009. Coalescent methods for estimating evolution of antibiotic resistance. Microbiol 77. Cui R, Schumer M, Kruesi K, Walter R, phylogenetic trees. Mol Phylogen Evol 53: Mol Biol Rev 74: 417–33.

et al. 2013. Phylogenomics reveals exten- 320–8. 109. Finley RL, Collignon P, Larsson DGJ, again Think sive reticulate evolution in Xiphophorus 93. Edwards SV. 2009. Is a new and general McEwen SA,etal. 2013. The scourge of fishes. Evolution 67: 2166–79. theory of molecular systematics emerging? antibiotic resistance: the important role of 78. Green RE, Krause J, Briggs AW, Mari- Evolution 63: 1–19. the environment. Clin Infect Dis 57: 704–10. cic T,etal. 2010. A draft sequence of the 94. Orr HA, Turelli M. 2001. The evolution of 110. Djogbenou L, Chandre F, Berthomieu A, Neandertal genome. Science 328: postzygotic isolation: accumulating Dobz- DabireR ,etal. 2008. Evidence of introgres- 710–22. hansky-Muller incompatibilities. Evolution sion of the ace-1R mutation and of the ace-1 79. Jarvis ED, Mirarab S, Aberer AJ, Li B,etal. 55: 1085–94. duplication in West African Anopheles 2014. Whole-genome analyses resolve early 95. Coyne JA, Orr HA. 2004. Speciation. gambiae s.s. PLoS ONE 3: e2172. branches in the tree of life of modern birds. Sunderland, MA: Sinauer Associates. 111. Weill M, Chandre F, Brengues C, Manguin Science 346: 1320–31. 96. Orr HA, Madden LD, Coyne JA, Goodwin S,etal. 2000. The kdr mutation occurs in the 80. Suh A, Smeds L, Ellegren H. 2015. The R,etal. 1997. The developmental genetics MoptiformofAnophelesgambiaes.s.through dynamics of incomplete lineage sorting of hybrid inviability: a mitotic defect introgression. Insect Mol Biol 9: 451–5. across the ancient adaptive radiation of in Drosophila hybrids. Genetics 145: 112. Norris LC, Main BJ, Lee Y, Collier TC,etal. neoavian birds. PLoS Biol 13: e1002224. 1031–40. 2015. Adaptive introgression in an African 81. Prum RO, Berv JS, Dornburg A, Field DJ, 97. Greig D. 2007. A screen for recessive malaria mosquito coincident with the in- et al. 2015. A comprehensive phylogeny of speciation genes expressed in the gametes creased usage of insecticide-treated bed birds (Aves) using targeted next-generation of F1 hybrid yeast. PLoS Genet 3: 281–6. nets. Proc Natl Acad Sci USA 112: 815–20. DNA sequencing. Nature 526: 569–73. 98. Greig D. 2008. Reproductive isolation in 113. Clarkson CS, Weetman D, Essandoh J, 82. Whelan NV, Kocot KM, Moroz LL, Hala- Saccharomyces. Heredity 102: 39–44. Yawson AE,etal. 2014. Adaptive introgres- nych KM. 2015. Error, signal, and the 99. True JR, Weir BS, Laurie CC. 1996. sion between Anopheles sibling species placement of Ctenophora sister to all other A genome-wide survey of hybrid incompati- eliminates a major genomic island but not animals. Proc Natl Acad Sci USA 112: bility factors by the introgression of marked reproductive isolation. Nat Commun 5: 4248. 5773–8. segments of Drosophila mauritiana chromo- 114. Adler PH, Cheke RA, Post RJ. 2010. 83. Hedrick PW. 2013. Adaptive introgression somes into Drosophila simulans. Genetics Evolution, epidemiology, and population in animals: examples and comparison to 142: 819–37. genetics of black flies (Diptera: Simuliidae). new mutation and standing variation as 100. Hollocher H, Wu CI. 1996. The genetics of Infect Genet Evol 10: 846–65. sources of adaptive variation. Mol Ecol 22: reproductive isolation in the Drosophila 115. Song Y, Endepols S, Klemann N, Richter 4606–18. simulans clade: X vs. autosomal effects D,etal. 2011. Adaptive introgression of 84. Mallet J. 2007. Hybrid speciation. Nature and male vs. female effects. Genetics 143: anticoagulant rodent poison resistance by 446: 279–83. 1243–55. hybridization between Old World mice. Curr 85. Rieseberg LH. 1997. Hybrid origins of plant 101. Masly JP, Presgraves DC. 2007. High- Biol 21: 1296–301. species. Annu Rev Ecol Syst 28: 359–89. resolution genome-wide dissection of the 116. Ellstrand NC, Meirmans P, Rong J, 86. Seehausen O. 2004. Hybridization and two rules of speciation in Drosophila. PLoS Bartsch D,etal. 2013. Introgression of adaptive radiation. Trends Ecol Evol 19: Biol 5: e243. crop alleles into wild or weedy populations. 198–207. 102. Wu CI, Palopoli MF, Johnson NA. 1996. Annu Rev Ecol Evol Syst 44: 325–45. 87. Stankowski S, Streisfeld MA. 2015. Intro- Haldane’s rule and its legacy: why are there 117. Alphey L. 2014. Genetic control of mosqui- gressive hybridization facilitates adaptive so many sterile males? Trends Ecol Evol 11: toes. Annu Rev Entomol 59: 205–24. divergence in a recent radiation of monkey- 281–4. 118. Hoffmann AA, Montgomery BL, Popovici flowers. Proc Biol Sci 282: 20151666. 103. Chapman T, Partridge L. 1996. Sexual J, Iturbe-Ormaetxe I,etal. 2011. Success- 88. Grant PR, Grant BR. 2014. Synergism of conflict as fuel for evolution. Nature 381: ful establishment of Wolbachia in Aedes natural selection and introgression in the 189–90. populations to suppress dengue transmis- origin of a new species. Am Naturalist 183: 104. Slotman M, della Torre A, Powell JR. sion. Nature 476: 454–7. 671–81. 2004. The genetics of inviability and male 119. Bian G, Joshi D, Dong Y, Lu P,etal. 2013. 89. Mallet J. 2013. Species, concepts of. In sterility in hybrids between Anopheles gam- Wolbachia invades Anopheles stephensi Levin SA, ed; Encyclopedia of Biodiversity, biae and An. arabiensis. Genetics 167: populations and induces refractoriness to Second Edition, Volume 6. Waltham, MA: 275–87. Plasmodium infection. Science 340: 748–51. Academic Press. p. 679–91. 105. Maynard Smith J. 1978. The Evolution of 120. Pike AD, Cirimotich CM, Dimopoulos G. 90. Behm JE, Ives AR, Boughman JW. 2010. Sex. Cambridge: Cambridge University 2013. Impact of transgenic immune deploy- Breakdown in postmating isolation and the Press. ment on mosquito fitness. In Koenraadt S, collapse of a species pair through hybrid- 106. Otto SP. 2009. The evolutionary enigma of Takken W, eds; Ecology and Control of ization. Am Naturalist 175: 11–26. sex. Am Naturalist 174: S1–S14. Vector-Borne Diseases, Volume 3, Ecology 91. Maddison WP. 1997. Gene trees in species 107. Fisher RA. 1930. The Genetical Theory of of Parasite-Vector Interactions. Wagenin- trees. Syst Biol 46: 523–36. Natural Selection. Oxford: Clarendon Press. gen: Academic Publishers. p. 19–33.

Bioessays 38: 140–149, ß 2015 The Authors. BioEssays Published by WILEY Periodicals, Inc. 149