SCIENCE ADVANCES | RESEARCH ARTICLE ANTHROPOLOGY Copyright © 2019 The Authors, some The comparative genomics and complex population rights reserved; exclusive licensee history of Papio baboons American Association for the Advancement of Science. No claim to Jeffrey Rogers1,2*, Muthuswamy Raveendran1, R. Alan Harris1,2, Thomas Mailund3, Kalle Leppälä3, original U.S. Government Georgios Athanasiadis3, Mikkel Heide Schierup3, Jade Cheng3, Kasper Munch3, Jerilyn A. Walker4, Works. Distributed Miriam K. Konkel5, Vallmer Jordan4, Cody J. Steely4, Thomas O. Beckstrom4, Christina Bergey6,7, under a Creative Andrew Burrell6, Dominik Schrempf8, Angela Noll9, Maximillian Kothe9, Gisela H. Kopp10,11,12, Commons Attribution Yue Liu1, Shwetha Murali1,13,14, Konstantinos Billis15, Fergal J. Martin15, Matthieu Muffato15, NonCommercial License 4.0 (CC BY-NC). Laura Cox16,17, James Else18, Todd Disotell6, Donna M. Muzny1,2, Jane Phillips-Conroy19,20, Bronwen Aken15, Evan E. Eichler13,14, Tomas Marques-Bonet21,22,23,24, Carolin Kosiol8,25, Mark A. Batzer4, Matthew W. Hahn26, Jenny Tung27,28,29, Dietmar Zinner10, Christian Roos9, Clifford J. Jolly6, Richard A. Gibbs1,2, Kim C. Worley1,2, Baboon Genome Analysis Consortium† Recent studies suggest that closely related species can accumulate substantial genetic and phenotypic differ- Downloaded from http://advances.sciencemag.org/ on February 10, 2019 ences despite ongoing gene flow, thus challenging traditional ideas regarding the genetics of speciation. Baboons (genus Papio) are Old World monkeys consisting of six readily distinguishable species. Baboon species hybridize in the wild, and prior data imply a complex history of differentiation and introgression. We produced a reference genome assembly for the olive baboon (Papio anubis) and whole-genome sequence data for all six extant species. We document multiple episodes of admixture and introgression during the radiation of Papio baboons, thus dem- onstrating their value as a model of complex evolutionary divergence, hybridization, and reticulation. These results help inform our understanding of similar cases, including modern humans, Neanderthals, Denisovans, and other ancient hominins. INTRODUCTION Baboons (order Primates; family Cercopithecidae; genus Papio) The increasing availability of genomic data across the tree of life has are large-bodied, geographically widespread Old World monkeys begun to challenge traditional concepts and assumptions regarding (OWMs; here, we use the term “baboon” to refer only to species in the genetics and population biology of phylogenetic differentiation the genus Papio). Their diversity provides an opportunity to investi- and speciation (1, 2). Reconstruction of the history of closely related gate the genomic, morphological, and behavioral aspects of an evolu- lineages suggests that cladogenesis (differentiation from a common an- tionary radiation in a broadly successful and adaptable primate. cestor that produces one or more new species) is often not as straight- Extensive work in both natural and captive populations has produced forward as assumed by traditional models of speciation (3, 4). Evolving considerable insight into baboon morphology, physiology, neuro- lineages may exchange functionally important genetic information biology, and behavior that provides additional context for compara- while remaining phenotypically distinct and diagnosable (5, 6). Com- tive evolutionary analyses (7–11). Mitochondrial DNA (mtDNA) paring genomic data across such closely related, divergent but still in- diversity suggests that this evolutionary radiation began about 2 mil- terfertile lineages provides new insight into cladogenesis in general and lion years (Ma) ago (12), approximately the same time as the radia- the nature, rate, and consequences of genomic evolution in particular. tion of our own genus, Homo, and in the same sub–Saharan African 1 Human Genome Sequencing Center, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA. 2Department of Molecular and Human Genetics, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA. 3Bioinformatics Research Centre, Aarhus University, CF Møllers Alle 8, DK-8000 Aarhus, Denmark. 4Depart- ment of Biological Sciences, 202 Life Sciences Building, Louisiana State University, Baton Rouge, LA 70803, USA. 5Department of Genetics and Biochemistry, 105 Collings Street, Clemson University, Clemson, SC 29634, USA. 6Department of Anthropology, New York University, 25 Waverly Place, New York, NY 10003, USA. 7Departments of Anthropology and Biology, Pennsylvania State University, 514 Carpenter Building, University Park, PA 16802, USA. 8Institut für Populationsgenetik, Veterinärmedizinische Universität Wien, Veterinärplatz 11210 Vienna, Austria. 9Primate Genetics Laboratory, German Primate Center, Leibniz Institute for Primate Research, Kellnerweg 4, 37077 Göttingen, Germany. 10Cognitive Ethology Laboratory, German Primate Center, Leibniz Institute for Primate Research, Kellnerweg 4, 37077 Göttingen, Germany. 11Depart- ment of Biology, University of Konstanz, Universitätsstr. 10, 78467 Konstanz, Germany. 12Department of Migration and Immuno-Ecology, Max Planck Institute for Ornithology, Am Obstberg 1, 78315 Radolfzell, Germany. 13Department of Genome Sciences, University of Washington, 3720 15th Avenue NE, S413C, Box 355065, Seattle, WA 98195-5065, USA. 14 Howard Hughes Medical Institute, University of Washington, 3720 15th Avenue NE, S413C, Box 355065, Seattle, WA 98195-5065, USA. 15European Molecular Biology Laboratory, European Bioinformatics Institute, Wellcome Genome Campus, Hinxton, Cambridgeshire, CB10 1SD, UK. 16Southwest National Primate Research Center, Texas Biomedical Re- search Institute, 8715 W. Military Drive, San Antonio, TX 78227, USA. 17Center for Precision Medicine, Department of Internal Medicine, Section on Molecular Medicine, Wake Forest School of Medicine, 475 Vine Street, Winston-Salem, NC 27101, USA. 18Department of Pathology and Laboratory Medicine and Yerkes Primate Research Center, 954 Gatewood Road, Emory University, Atlanta, GA 30322, USA. 19Department of Neuroscience, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110, USA. 20Department of Anthropology, Washington University, McMillan Hall, 1 Brookings Drive, St. Louis, MO 63130, USA. 21Institute of Evolutionary Biology (UPF-CSIC), PRBB, Dr. Aiguader, 88. 08003, Barcelona, Spain. 22Catalan Institution of Research and Advanced Studies (ICREA), Passeig de Lluís Companys, 23, 08010, Barcelona, Spain. 23CNAG- CRG, Centre for Genomic Regulation (CRG), Barcelona Institute of Science and Technology, Baldiri Reixac, 4, 08028, Barcelona, Spain. 24Institut Catala de Paleontologia Miquel Crusafont, Universitat Autonoma de Barcelona, c/de les Columnes, s/n. Campus de la UAB. 08193–Cerdanyola del Vallès, Barcelona, Spain. 25Centre for Biological Diversity, School of Biology, University of St. Andrews, Dyers Brae House, Greenside Place, St Andrews, Fife, KY16 9TH, UK. 26Department of Biology and Department of Computer Science, Indiana University, 1001 E. 3rd Street, Bloomington, IN 47405, USA. 27Department of Biology, Duke University, Box 90338, Durham, NC 27708, USA. 28Duke Population Research Institute, Duke University, Box 90989, Durham, NC 27708, USA. 29Institute of Primate Research, P.O. Box 24481, Nairobi, Kenya. *Corresponding author. Email: jr13@bcm.edu †For a complete list of members of the consortium, see the Acknowledgments section. Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 1 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE environment (13). However, unlike Homo, which is now reduced to a These results, for a clade in which multiple lineages and active hybrid single surviving species (Homo sapiens), Papio still includes six extant zones are available for direct observation, provide important insights lineages, the products of successive speciation events. These six spe- relevant to other cases of complex differentiation and admixture, in- cies are morphologically and behaviorally distinct (10–12) and have cluding that of ancestral modern humans and our extinct relatives broad, adjoining but nonoverlapping geographic ranges across sub– such as Neanderthals, Denisovans, and other early human lineages. Saharan Africa and southwest Arabia (Fig. 1; see Supplementary Text for explanation of the taxonomy used here). The morphological and behavioral traits that define each species are unambiguous (Fig. 1) RESULTS and expressed quite homogeneously over large geographic distances We produced a whole-genome reference assembly (Panu_3.0; GenBank (10, 11, 14). accession GCA_000264685.2) for an olive baboon (Papio anubis), the Despite these phenotypic differences, genetic evidence reveals an baboon species most commonly used in biomedical research (fig. S1 underlying complexity to baboon evolutionary history. Maternally and tables S1 and S2) (7). To investigate genetic differentiation in the inherited mtDNA yields a phylogeny that includes at least seven ma- genus, we analyzed whole-genome sequences from 16 additional in- jor haplogroups whose distribution is discordant with the relation- dividuals, 2 to 4 individuals representing each of the six species within ships implied by phenotypic comparisons (12, 15). Since migration Papio, and a gelada (Theropithecus gelada), a member of a closely re- among baboon social groups and populations is generally sex-biased, lated genus that serves as an outgroup (fig. S2 and table S3). This diver- with males usually (but not always) the dispersing sex (16), popula- sity panel produced >54.6 million single-nucleotide variants (SNVs), tion relationships based on maternally inherited mtDNA will not of which >42.4 million are variable within Papio (fig. S3 and table S4). Downloaded from http://advances.sciencemag.org/ on February 10, 2019 necessarily correspond to population relationships based on nuclear To develop a second independent perspective on genome differentia- DNA or phenotype. Furthermore, baboon species produce fertile tion, we identified novel Alu insertions, a type of genetic variation that hybrid offspring in the wild and can form long-lasting hybrid zones results from a fundamentally different mutational mechanism. Unex- (17–19) despite substantial species-specific differences in body size, pectedly, we found a dramatically elevated number of recent Alu inser- secondary sexual characteristics, and social systems (11). tions in the baboons (and in rhesus macaques) relative to human and This constellation of genetic and phenotypic observations raises a other primate genomes (Fig. 2 and table S5). There are 192,889 full- number of questions regarding the history of differentiation and ge- length AluY elements in the P. anubis genome. The rate of accumula- netic exchange among baboons. Is baboon hybridization a recent tion of lineage-specific AluY insertions has therefore been more than phenomenon, or does the discordance between the mtDNA phylog- fourfold higher (Fig. 2) in baboons and rhesus macaques than in homi- eny and phenotypic similarity result in part from ancient episodes of noids (humans, chimpanzees, or orangutans) and about threefold admixture and evolutionary reticulation? Has admixture been lim- higher than in the African green monkey (genus Chlorocebus), an- ited to the extant lineages within Papio, or as has been suggested for other OWM (22). hominins (20, 21), have now extinct “ghost” lineages also played a role? Our phylogenetic analyses provide several new insights into ba- The current study investigates baboon genomic diversity to provide a boon population and genomic history. Maximum likelihood (ML) detailed reconstruction of the history of evolutionary differentiation analyses of concatenated SNVs show that individual baboons cluster among the six extant species and new insights regarding the pro- correctly with their conspecifics while separating the six extant spe- cesses, correlates, and consequences of genomic differentiation. cies into distinct northern and southern clades (Fig. 3 and fig. S4A). Fig. 1. Papio baboon species. (A) The appearance and current distribution of each baboon species, and the locations of three well-documented active hybrid zones are also shown. x1: hybrid zone between P. hamadryas and P. anubis (19, 28), x2: hybrid zone between P. cynocephalus and P. anubis (17, 26), x3: hybrid zone between P. kindae and P. ursinus (18). Drawings of each species by S. Nash. (B) Distinguishing features of Papio species. Body mass data from (16, 59) and unpublished data from J.P.-C., J.R., and C.J.J. Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 2 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE In contrast, Bayesian analysis of the same SNV data suggests that affect inferred branch lengths and that lineages that have experienced P. kindae is sister to the northern clade rather than to P. cynocephalus admixture will exhibit artificially shorter branch lengths due to allele and P. ursinus (fig. S4B). The existence of multiple hybrid zones and sharing across lineages. This suggests that the other four lineages may documented discrepancies between relationships based on mtDNA have been more affected by admixture than P. ursinus and P. papio, and on phenotypes (Fig. 1) (12, 15) argue that male-driven admixture which is consistent with the fact that these two species are found at and/or incomplete lineage sorting (ILS) have influenced genetic rela- the extreme southern and western reaches of baboon distributions, re- tionships among these species. When we used a polymorphism-aware spectively (Fig. 1). phylogenetic approach, PoMo (23, 24), we again obtained a basal north- To explicitly test for admixture among the six extant baboon spe- south divergence, with P. kindae placed in the southern group. How- cies, we performed an analysis using f-statistics, followed by model- ever, the relationships among the three southern species differ from ing using coalescent hidden Markov methods (Fig. 4A, table S7, and the ML result (Fig. 3). PoMo also infers much longer terminal branch fig.S6). The best-fitting model (see Materials and Methods) indicates lengths for P. ursinus and P. papio than for other lineages. Simulations that the history of P. kindae includes an ancient admixture event involving (fig. S5 and table S6) show that admixture among divergent lineages can a lineage related to extant P. ursinus (52% contribution to extant P. kindae) and an unsampled lineage (possibly extinct) belonging to the northern clade (48% contribution). The f-statistics suggest that extant P. papio is closely related to P. anubis, but received ~10% genetic input from an ancestral northern lineage also not yet sampled, possibly extinct. Our results also shed new light on the historical dynamics of hybrid- Downloaded from http://advances.sciencemag.org/ on February 10, 2019 ization between P. anubis (a northern clade species) and P. cynocephalus (a southern clade species), which has previously been reported in south- ern Kenya near Amboseli National Park (17). Behavioral observations and microsatellite-based analyses support recent introgression from P. anubis into P. cynocephalus since the 1980s (25, 26). Our analysis of genome-wide haplotype block sharing indicates that a P. anubis individual from the Aberdare region of Kenya, more than 200 km north of Amboseli, is also admixed with P. cynocephalus, carrying ~546 Mb of nuclear DNA derived from P. cynocephalus (fig. S7). If we assume that this resulted from a single admixture event, then it is estimated to have occurred about 21 generations (~220 years) ago. However, other more complex explanations are also possible. The second individual from the P. anubis Aberdare population also carries Fig. 2. Comparison of Alu mobilization rates in selected primate genomes. Only P. cynocephalus haplotypes, but these shared genomic segments are Alu elements specific to each lineage are included. The size of the circle corresponds to fewer and shorter and likely result from more ancient introgression. the number of near full-length lineage-specific AluY elements in that species. The bars Consistent with other studies (27), our findings suggest that there have on the right show the estimated number of insertions per million years for each been multiple episodes of gene flow involving these two species over a lineage. For baboon (Panu_3.0), rhesus macaque (Mmul_8.0.1), African green mon- considerable time span and that the effects of past hybridization ex- key (chlSab2), chimpanzee (Pan_tro3), and human (GRCh38/hg38), AluY sequences were retrieved computationally by cross comparisons using the most recent avail- tend far beyond the current hybrid zone. This complexity may well able assemblies. Orangutan estimates are from P_pygmaeus2.0.2 (60). The number be representative of the complexity of other known baboon hybrid of lineage-specific AluY elements is similar in rhesus macaque and baboon, and zones (10, 12, 15, 18, 19, 28). more than twice that in the African green monkeys, despite a longer period of Motivated by the results from the f-statistics and haplotype sharing, independent evolution for the African green monkeys. we conducted two additional tests across the Papio diversity panel to Fig. 3. Phylogenetic relationships among baboon species. (A) Phylogeny generated using the polymorphism-aware phylogenetic method (PoMo) (23, 24). This topology for the three northern species is also supported by ML analysis of concatenated SNVs and by 43.9% of informative gene trees filtered to exclude any coding sequence genes [scaled concordance factor (CF) of 0.439, greater than the other two alternatives]. The topology shown for the three southern clade species is supported by the PoMo analysis and has a scaled CF score of 0.332. (B) One alternative topology for the northern species, supported by a scaled CF of 0.241. (C) One alternative topology for the southern species, supported by ML analysis of concatenated SNVs and a scaled CF score of 0.513, i.e., a larger proportion of gene trees that are devoid of coding genes than the other two alternative trees. Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 3 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE Downloaded from http://advances.sciencemag.org/ on February 10, 2019 Fig. 4. Evolutionary and demographic history for Papio baboons. (A) Analyses using f-statistics indicate that P. kindae was formed via input from both a southern clade lineage and a northern clade lineage, with contributions estimated to be 52 and 48%. P. papio is inferred to have been produced through 10% introgression from an unidentified ancient northern lineage into a population related to P. anubis. Dates for divergence and admixture events were inferred through CoalHMM, and internal nodes representing those divergence or admixture events are labeled A through K. Our analyses of asymmetric haplotype sharing also inferred admixture from P. cynocephalus into P. anubis approximately 21 generations ago. (B) Reconstruction of baboon demographic history using PSMC methods. A prolonged bottleneck was observed in the lineage ancestral to P. papio beginning ~400 thousand years (ka) ago, while the populations ancestral to P. hamadryas and P. anubis increased between ~280 and ~160 ka ago. After diverging, P. anubis followed an upward trend whereas P. hamadryas declined. At ~400 ka ago, Ne for P. ursinus diverged from estimates for the populations ancestral to P. cynocephalus and P. kindae, and underwent a species-specific prolonged bottleneck. At ~300 ka ago, the Ne reconstructed for both P. cynocephalus and P. kindae increased, peaking ~150 ka ago before experiencing a subsequent decline. PSMC methods are not always reliable for the most recent time periods. Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 4 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE examine the hypothesis of ancient admixture using independent conclusion that all baboon species share the same demographic history methods. Alu insertion polymorphisms are valuable phylogenetic (that is, were effectively one lineage) before ~1.4 Ma ago. All Ne plots characters because the polarity of specific mutational changes can be show an upward trend after ~1.5 Ma ago, but the species-specific in- unambiguously established for any given genomic segment (fig. S8) creases occur at different rates, possibly corresponding to population (29). The haplotype that carries a novel Alu insertion is derived from growth and dispersion once ecological conditions allowed demographic the orthologous haplotype lacking the Alu repeat, and reversals are rare. expansion (14). Given the paleontological evidence for a southern origin Majority-rule Dollo parsimony analyses of the baboons using novel Alu of this genus (36), we speculate that the more pronounced apparent de- insertions once again revealed a north-south difference. However, the cline in Ne for northern clade species relative to the southern lineages descendent lineages are poorly resolved, exhibiting apparent homoplasy ~700,000 to 800,000 years ago may reflect dispersal-related bottlenecks (fig. S8). In a phylogeny constructed using characters with well-defined as the geographic range of baboons was extended to the north. Similarly, polarity, such homoplasy would not be expected unless a radiation of the CoalHMM suggests that the north-south admixture that produced species experienced substantial ILS and/or gene flow among divergent extant P. kindae occurred about 100,000 years ago, and the PSMC re- lineages (30). sults suggest an increase in Ne for P. kindae about this time. We next examined differences in the evolutionary history of dif- To examine potential functional consequences of baboon admix- ferent segments across the baboon genome. We divided the reference ture, we investigated 2201 suitable genic regions (local genomic seg- genome into 808 discrete gene-free (putatively neutral) regions. ments that contain one annotated protein-coding gene each and Using BUCKy (31) and the SNV genotypes from the diversity panel, exhibit sufficient phylogenetic signal to support one particular phy- we performed Bayesian concordance analysis (BCA). Individual logenetic tree over all alternative trees). We identified individual loci Downloaded from http://advances.sciencemag.org/ on February 10, 2019 animals again, as expected, cluster by species. The basal north-south that exhibit phylogenetic relationships (gene trees) concordant or divergence is supported, but the concordance factors (CFs) for rela- discordant with the consensus species-level phylogeny that sep- tionships within each of those two geographic clades are low (Fig. 3). arates the three northern species from the three southern species. P. hamadryas is most frequently sister to an anubis-papio clade, but Cluster 1 contains 1143 genic regions with phylogenies closely match- the two other possible topologies [(papio-(ham-anubis)) and (anubis- ing that result (fig. S16). Cluster 2 consists of 629 genic regions for (ham-papio))] do not appear at equal frequency (fig. S9), as would be which P. cynocephalus carries haplotypes that are not closely related expected under ILS. Similarly, P. kindae is most frequently sister to a to other southern clade haplotypes (fig. S17). The genes in these re- cynocephalus-ursinus clade, matching the ML results but not f-statistics gions are enriched for the gene ontology (GO) terms “learning and or PoMo results. Again, the two minor BCA topologies are not found in memory” (P = 0.012), “cognition” (P = 0.012), “head development” equal proportions (fig. S9). Together, the Alu insertion and BCA results (P = 0.014), and “brain development” (P = 0.017), as well as several support the conclusion that reticulation rather than ILS without retic- GO categories related to reproduction (see table S10). Cluster 3 in- ulation has influenced baboon genomic divergence (Table 1). cludes 429 genic regions displaying phylogenetic relationships The timing of lineage divergences and admixture events was es- among southern clade species consistent with the phylogeny in timated using a coalescent hidden Markov model (CoalHMM; figs. Fig. 3a. However, the cluster 3 haplotypes from the northern clade S10 to S15 and tables S8 and S9) (32, 33). Using an estimated muta- P. anubis are more closely related to haplotypes from the southern tion rate of 0.9 × 10-8 per base pair per generation and a generation clade, while haplotypes in northern clade P. papio generally form the time of 11 years [see Materials and Methods and (11, 34)], we obtain sister to all other baboon haplotypes (fig. S18). The genes found in the results presented in Fig. 4A. To reconstruct demographic history, cluster 3 regions are enriched for GO terms related to the ontogenet- we generated pairwise sequential Markovian coalescent (PSMC) ic development of several organ systems (kidney, heart, circulatory, plots (35), assuming the generation time and mutation rate cited and endocrine systems, all significantly enriched with P < 0.03) (table above (Fig. 4B). With the exception of P. papio, which has a truncated S10). We note that the two species exhibiting the clearest gene tree plot, the remaining five species are very similar in effective popula- discrepancies relative to the species-level phylogeny (i.e., species carry- tion size (Ne) from 4 Ma ago up until ~1.4 Ma ago, supporting the ing haplotypes that apparently crossed species boundaries) are P. anubis Table 1. Summary of diverse data types and analytical approaches used to investigate the phylogeny of baboon species. Data type Analytical method Primary clustering Northern clade result Southern clade result SNVs ML phylogeny North versus south hamad-[papio-anubis] kinda-[cyno-ursinus] SNVs Bayesian phylogeny North plus kinda versus kinda-[hamad-[papio-anubis]] [cyno-ursinus] two south SNVs Polymorphism-aware phylogeny North versus south hamad-[papio-anubis] cyno-[kinda-ursinus] SNVs f-statistics North versus south hamad-[papio-anubis] cyno-[kinda-ursinus] Alu insertion polymorphisms Majority-rule Dollo parsimony North versus south Unresolved phylogeny with Unresolved phylogeny with north-south split north-south split Locus-specific trees for segments Concordance factors North versus south hamad-[papio-anubis] kinda-[cyno-ursinus] without coding sequences Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 5 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE and P. cynocephalus, a northern clade and a southern clade species, of thousands of years (20, 21). Baboons have such a history but can respectively, that actively hybridize today in southern Kenya (17) still be studied in present-day hybrid zones and therefore constitute and exhibit evidence of nuclear DNA swamping (12). an important context for future research. Potential areas of study in- clude the effects of genetic variation on neurotransmitter function and its impact on species-level differences in social relationships DISCUSSION and social behavior (41, 42). The presence of hybrid zones between Like our own genus Homo, the ancestral stock of Papio baboons began species pairs separated by different genetic distances (e.g., the distant diverging into multiple lineages within sub–Saharan Africa by about P. anubis/P. cynocephalus versus the much closer P. anubis/P. hamadryas) 1.5 Ma ago. Like early Homo, baboon species today differ in body size makes it feasible to investigate the effects of increasing genetic differen- and morphology (9, 11, 13). We report here that multiple baboon tiation. Access to well-characterized captive research colonies of ba- lineages have experienced episodes of admixture, some involving ge- boons provides further opportunity for innovative studies concerning netic exchange among lineages that persist today while other episodes developmental, metabolic, and neurobiological consequences of inter- involved extinct lineages. In contrast to Homo, Papio today includes six breeding among divergent lineages (7, 39, 43, 44). surviving differentiated lineages (phylogenetic species), providing a unique context for the investigation of genetic and phenotypic consequences of both ancient and modern interspecies hybridization. MATERIALS AND METHODS No one simple dichotomously branching tree accurately reflects Sequencing of the reference genome sample all aspects of genomic differentiation among extant baboon species. The sequence data for the olive baboon (P. anubis) whole-genome as- Downloaded from http://advances.sciencemag.org/ on February 10, 2019 However, our provisional scenario for baboon genome evolution, sembly were generated through different methods over time, as the presented in Fig. 4A, does establish the context for further explora- dominant sequencing technologies evolved. The earliest data were gen- tions of baboon biology. We observed a markedly increased rate of erated using the Sanger technology, followed by Roche 454 FLX data. recent Alu insertion mobilization in the baboons relative to human Later, Illumina short-read data (both paired-end and mate-pair and other hominoids (Fig. 2). Previous studies suggest that hybrid- reads) were generated using the Genome Analyzer IIx first and then ization between divergent lineages can generate increases in the rate the HiSeq 2000 platforms later. Last, Pacific Biosciences RSII data of novel insertions of repetitive elements (37, 38). Another topic of were also produced. All the Sanger, Roche 454, and Illumina read data broad interest is the origin of reproductive isolation among incipient were generated from a single female olive baboon of Kenyan ancestry species (1). One expectation for the genus Papio is that, given the timing [animal ID 1X1155, National Center for Biotechnology Information of the radiation and the degree of morphological and behavioral differ- (NCBI) BioSample SAMN02981400; Southwest National Primate Re- entiation among species, incipient barriers to gene flow may be evident search Center, San Antonio, TX]. The PacBio read data were generated between some pairs of species. Studies of the present-day hybrid zone from a single olive baboon (animal ID 20111; assigned three BioSample between northern clade P. anubis and southern clade P. cynocephalus numbers: SAMN03165174, SAMN03165175, and SAMN03165176) find no readily apparent barriers to reproduction between these species from the same research colony. The depth of genome coverage used (17, 26). However, studies of captive P. anubis × P. cynocephalus hybrids in the assembly was as follows: Sanger 2.5×, Roche 454 4.5×, Illumina document significantly elevated frequencies of craniodental anomalies 85×, and PacBio RSII 12×. All reads have been deposited in NCBI under in hybrids, especially hybrid males, indicating some degree of genetic BioProject PRJNA54005. incompatibility (39). Field studies of the hybrid zone between P. ursinus and P. kindae describe a deficit of hybrid individuals carrying Y chro- Genome assembly mosomes from P. ursinus and mtDNA from P. kindae compared to the The genome assembly processes used are shown in fig. S1. The initial olive converse (18). This suggests that when hybridization began between baboon genome assembly, Pham_1.0, used only Sanger and Roche 454 these two forms, some type of barrier (premating or postmating) re- data. This assembly is no longer available at NCBI but can still be accessed duced the frequency or fertility of matings by male P. ursinus with fe- at the University of California, Santa Cruz (UCSC). To avoid confusion, male P. kindae, while the converse mating type was more successful we emphasize that although listed under Baboon (hamadryas) and (18). Last, P. anubis and P. hamadryas differ substantially in their social named papHam1 on the USCS genome browser and at the Ensembl organization and social structure (11, 28, 40). Among anubis baboons, Pre! Site (http://pre.ensembl.org/Papio_hamadryas/Info/Index), this first both males and females are polygamous. Hamadryas societies are multi- version of the assembly was derived not from a hamadryas baboon but level, with “harem”-like, one-male breeding units (OMUs) as basal so- from the female olive baboon identified above. The analyses reported here cial entities. In these OMUs, the single adult male defends exclusive only used the later improved assemblies, Panu_2.0 and Panu_3.0. access to one or more adult females. Other differences in sex-specific Panu_2.0 (named Panu_2.0 in NCBI and papAnu2.0 in Ensembl dispersal and social relationships are also observed (11). Despite the dra- and UCSC; GenBank accession GCA_000264685.1) was produced matic differences in social systems, these species hybridize in the wild from the available Sanger, Roche 454, and Illumina reads, derived from (28). Hybrid males can achieve substantial reproductive success, at least the same female olive baboon used for Pham_1.0. Assembly analyses in groups consisting mainly of hybrids (19). There is no clear evidence used the GAC (Genomic Analysis Cluster) compute facilities at the for a barrier to gene flow between the species, although the geographic Baylor College of Medicine Human Genome Sequencing Center distribution of phenotypically recognizable hybrids is narrow. (HGSC). Sanger and Roche 454 reads were first assembled using The demonstration that hybridization among modern humans, CABOG version 6.1 with parameter settings of utgErrorRate = Neanderthals, and Denisovans had enduring effects on the modern 0.02, ovlErrorRate = 0.07, cnsErrorRate = 0.07, cgwErrorRate = 0.12, human gene pool has raised questions about the demographic pro- and unitigger = bog. Two sets of 100–base pair (bp) Illumina read data, cesses, as well as the genomic and phenotypic consequences, of ad- 2 billion reads from a 240-bp insert paired-end library, and 500 million mixture among primate lineages separated on the order of hundreds reads from a 2.5-kb insert mate-pair library were mapped to the Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 6 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE CABOG assembly using BWA with default parameters. The scaffolds of obtained from captive colonies. The species identity of each sampled this initial CABOG-generated assembly were improved on the basis of animal was determined from its external phenotype. The integrity the read mapping locations using Atlas-Link version 1.0 (https://www. of subsequent sequence data files was confirmed by comparing the hgsc.bcm.edu/software), with the minimum required links (min_link) mtDNA sequences obtained through whole-genome analysis to other set at four for the 240-bp library and three for the 2.5-kb library. The mtDNA sequences from baboons of known species and geographic lo- Atlas-GapFill version 1.0 process (https://www.hgsc.bcm.edu/ cation (12). All such species assignments were confirmed and validated. software) was then performed to fill gaps between contigs within All these diversity samples were sequenced to an average read depth of scaffolds by extracting local read pairs and aligning the local assem- 30.7× using the Illumina HiSeq 2000 sequencing platform (100-bp blies of these pairs to the gaps. paired-end reads), with the one exception that the T. gelada sample The assembled contigs and scaffolds of Panu_2.0 were placed on ba- was sequenced on the Illumina HiSeq X platform. boon chromosomes by mapping to the rhesus macaque (Macaca We used BWA-MEM version 0.7.12-r1039 (https://arxiv.org/abs/1303. mulatta) genome assembly (GCF_000002255.3, mmul_051212, rhemac2) 3997) to align the Illumina reads to the baboon reference assembly Panu3.0/ using Mummer3 (parameters = nucmer -l 12 -c 65 -g 1000 -b 1000; delta- papAnu3 and generate BAM (Binary Alignment Map) files (fig S2). Picard filter -1 -l 500; show-coords -cl -L 500). It should be noted that chro- MarkDuplicates version 1.105 (http://broadinstitute.github.io/picard/) was mosome organization is largely conserved between rhesus macaque and used to identify and mark duplicate reads. Variants were called using GATK baboon (45). A baboon scaffold was split when it did not have contin- version 3.3-0 following best practices for that version (https://software. uous alignment on the macaque genome and if the potential breakpoint broadinstitute.org/gatk/best-practices/). In brief, indels were realigned was validated by low clone coverage in the baboon data (low cover- using IndelRealigner. HaplotypeCaller was used to generate gVCFs for Downloaded from http://advances.sciencemag.org/ on February 10, 2019 age defined as clone coverage from the 2.5-kb Illumina library of <5×). each sample. Joint genotype calling was performed on all samples using A set of 323 scaffolds (a total of 217 Mb) were identified this way and GenotypeGVCFs to generate a VCF file. GATK hard filters (SNPs: “QD < therefore split. The N50 of the contigs in the Panu_2.0 assembly is 2.0 || FS > 60.0 || MQ < 40.0 || MQRankSum < −12.5 || ReadPosRankSum < 40.3 kb, and the N50 of the scaffolds is 529 kb. The total length of the −8.0”; Indels: “QD < 2.0 || FS > 200.0 || ReadPosRankSum < −20.0”) (https:// Panu_2.0 assembly is 2.95 Gb with 55.1 Mb of gaps. Because the scaffolds software.broadinstitute.org/gatk/documentation/article?id=2806) were for Panu_2.0 (and Panu_3.0) have been mapped onto baboon chromo- applied, and all variant calls that failed the filters were removed. somes, this genome assembly is presented in public databases (NCBI, To perform functional annotations through WGSA (Whole Genome UCSC, and Ensembl genome browsers) as chromosome-associated se- Sequence Annotator) (48), the SNVs identified in the baboon diversity pan- quences rather than as sets of independent scaffolds and superscaffolds. el were transferred to the human genome (hg19) using liftOver and treated Last, we improved the Panu_2.0 assembly through two additional as human SNVs. All annotation resources available for version 0.5 were methods. First, a small number of differences between the baboon used for this analysis, including five functional prediction scores, eight con- and rhesus macaque genomes were identified using fluorescence in situ servation scores, allele frequencies from four large-scale resequencing hybridization (FISH) mapping of probes containing human BAC se- studies, and variants in four disease-related databases, among others. quences. These scaffolds were refined to be consistent with the FISH results from the baboon genome. Last, a total of 12× whole-genome cov- Alternative phylogenetic analyses erage was produced on the PacBio RSII platform, with half of the reads Given the clear discordances between mitochondrion-based and phenotype- >7 kb. These data were mapped to the Panu_2.0 assembly and two- based phylogenies (12, 15), we performed an extensive series of phylo- thirds (67%) of the 118,928 gaps within scaffolds were closed using genetic analyses using different data types and analytical approaches. PBJelly software (46). The base quality of the assembly was polished Our goal was to develop robust conclusions regarding population using the Pilon program (47) and the available Illumina data. history that are supported by multiple datasets and analyses. This final assembly (Panu_3.0) has a contig N50 of 149.8 kb and, Phylogenetic analysis of concatenated whole-genome due to the mapping of these scaffolds to chromosomes, it has near whole SNV data chromosome length superscaffolds. The gap filling with PBJelly added The dataset for this analysis consisted of SNV calls for the 15 baboons only 10.98 Mb to the assembly (0.37% of the Panu_2.0 assembly length), from the diversity panel (table S4), plus one gelada (T. gelada). From the but closed a large number of gaps, reducing the number of contigs from variant call file (vcf) produced by GATK, only SNV positions surviving 198,931 to 118,251. The Panu_3.0 assembly was tested against available filtering steps conducted with vcffilter from vcflib (https://github.com/ baboon EST (Expressed Sequence Tag) sequence datasets to quantify ex- ekg/vcflib; settings: -f "QUAL > 20 & DP > 10 & MQ > 30 & QD > 20") tent of coverage (i.e., completeness). Of the 144,708 Sanger EST sequences were used for further analysis (24,588,548 SNVs). SNVs were extracted available at the time of testing, 99.98% were successfully mapped to the from the filtered vcf file with bcftools from SAMtools 1.2 (settings: bcftools assembly. Among the total ESTs, 98.77% mapped with >90% of their query -f ’%CHROM\t%POS\t%REF\t%ALT [\t%SAMPLE =%GT]\n’). length and 97.48% mapped at >95% of length. Seven finished BAC clones The resulting table was converted into individual FASTA sequences were mapped to the Panu_2.0 assembly. The genomic coverage in the using a custom Python script. Individual cases where a baboon exhib- BACs was high, with 98 to 100% of the BAC sequence in the assembly. ited more than one different nonreference allele at the same site were The assembled contigs and scaffolds were aligned linearly to the finished recorded as ambiguous. Merging all FASTA sequences into a single file BACs, suggesting that misassemblies are rare. Within Panu_3.0, only provided a multiple sequence alignment of all individuals and all con- 3.2% of the sequence falls in unscaffolded contigs. catenated SNVs. Positions in the alignment where no information was given for at least one species were removed. A total of 22,433,604 SNVs Sequence variation across the diversity panel remained for analysis. Model selection using the Bayesian information DNA was obtained from 16 animals representing all six species of Papio criterion in IQ-TREE 1.3.13 (49) revealed the TVM+ASC+G model as the baboons and the gelada, T. gelada (table S3). Of the 16 individuals, 9 best-fit model for this dataset. Phylogenetic trees were reconstructed were wild animals sampled in the field, and the remaining samples were with ML and Bayesian approaches using IQ-TREE and MrBayes 3.2.6 Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 7 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE (50), respectively. IQ-TREE settings: TVM+ASC+G model, 1000 ultra- Z), where W was rhesus macaque and X, Y, and Z were all combinations fast bootstraps; MrBayes settings: TVM+G, 100,000 generations and of baboon samples from three different species. The bulk of our 10% burnin. analyses of f-statistics used sequence diversity and SNV data based Polymorphism-aware phylogenetic model analysis on mapping reads to the Panu_2.0 genome assembly. Once the im- of whole-genome data proved Panu_3.0 assembly was complete, which closed thousands of To estimate species-level phylogeny while allowing for current and pos- gaps in scaffolds but increased the total sequence length of the sible ancient polymorphism, we applied the PoMo model (23) imple- assembly by only 0.37%, we retested the likelihoods of the inferred mented in IQ-TREE (49) to the baboon diversity Panu_2.0 SNV data phylogenetic relationships among lineages and the inferred history together with fourfold degenerate sites of the orthologous gene set. of admixture using SNVs called based on Panu_3.0. All conclusions Briefly, the PoMo model represents the evolution of an individual nu- regarding species phylogeny and admixture events that were based cleotide site within a given fixed species-level phylogeny as a continuous on Panu_2.0 data were confirmed when tested using SNV and diver- time Markov chain along that phylogeny. Rather than considering only sity information based on read mapping against Panu_3.0. This is four states (four alternative nucleotides) for a given genomic position, likely because the upgrade from Panu_2.0 to Panu_3.0 added little PoMo allows for polymorphism within species by expanding the state to the total sequence length. space in the Markov chain to include heterozygous nucleotide compo- Grouping of samples sitions, assuming two nucleotides per site, in addition to the traditional When estimating f-statistics, we can either pool samples from the same four nucleotide states. Mutation (e.g., using the HKY model) introduces population together or compute at the level of individual samples. Pool- new nucleotides. The Moran model was used to describe genetic drift or ing samples would potentially give a better estimate of population allele Downloaded from http://advances.sciencemag.org/ on February 10, 2019 changes in allele frequencies over time. PoMo generates a single species frequencies, but can mask within-population differences that might be tree, but does allow for ILS. Additional details are available in (23, 24). informative about recent gene flow. We therefore chose to estimate the Simulation study comparing methods statistics at the individual sample level. For this analysis, we estimated We analyzed the robustness of the PoMo results to admixture between the f4-statistics for all triplets of baboon samples combined with rhesus differentiating lineages using the baboon phylogeny as the assumed con- macaque. These were computed using the qpDstats tool from the text. We defined the input phylogeny as that obtained by modeling ADMIXTOOLS package (https://github.com/DReichLab/AdmixTools). potential admixture among the six baboon species through f-statistics Inferring admixture graphs with each sample as a leaf is computationally (see below). Total branch lengths for each lineage were set as inferred from intractable. The number of possible graphs grows superexponentially baboon data, and Watterson’s q within species was set to 0.0025. We tested with the number of leaves, and our brute-force approach to exploring the ability of PoMo to accurately reconstruct the phylogeny for P. kindae the graph space only scales to a small number of leaves. We therefore by varying the proportion of admixture into the P. kindae lineage from a chose to keep the graphs at the species level. This way, samples from northern clade species from 0 (no admixture) to 80%. We simulated the same species are expected to have the same relationship to all other 1000 genes (1000 bp per gene) on five chromosomes and created species. Parameters are estimated from the combined set of individual 1000 gene trees using MSMS (51). We next concatenated the sequence samples within each species. data for five chromosomes and all gene trees in each species. We then We inspected the estimated f-statistics to ensure that the species used both PoMo and the HKY model to generate phylogenies and com- grouping matched with similar vectors of statistics. In general, samples pared their ability to reconstruct the correct species-level phylogeny. from the same species had very similar f-statistics compared to all other samples. The single exception was P. anubis sample 30877 from the Analyses of admixture through f-statistics Aberdare region of Kenya that has a different profile from the other Admixture graphs (52) model the ancestry of a set of samples in the olive baboons. This sample shows evidence of a recent admixture with form of a directed acyclic graph where edges capture drift along ances- P. cynocephalus. To focus on the ancient admixture events only, we re- tral lineages, leaves represent the samples, and inner nodes represent moved this sample from the admixture graph analysis; thus, the species either most recent common ancestral populations or admixture events “anubis” in the following refers only to the remaining P. anubis samples. where a new population is created as a mixture of two other popula- Fitting admixture graphs tions. Hence, these graphs can capture more complex histories than A given graph topology specifies a polynomial of edge lengths and admix- simple tree phylogenies, but only simple forms of gene flow. Admixture ture proportions as the expected value for each f-statistic. These poly- graphs model all gene flow as admixture events and cannot easily model nomials are linear with respect to the edge lengths only, and so we periods of continuous gene flow. Admixture graphs are parameterized stored them as rows in a matrix of polynomials of admixture proportions by edge lengths (the amount of drift that occurred on a given ancestral only. To measure the fit between a graph and the observed statistics, we lineage) and admixture proportions (how much of an ancestral ad- defined a cost function: a weighted sum of squared errors between graph mixed population was derived from one donor population rather than predictions F and statistics f. The weights are reciprocals of the SDs of the another). Properties of an admixture graph can be captured by so-called statistics f, given by ADMIXTOOLS (as the Z values divided by f ). We f-statistics (52), and these can be estimated from genomic data. This fitted the parameters of the graph using a mix of analytic and numerical makes it possible to compare the statistics predicted by a graph, F, with optimization. After fixing the admixture proportions, the polynomial statistics estimated from data, f. Graph parameters are estimated by equations of predictions F are linear and thus solvable analytically. To minimizing the distance between F and f. We have implemented an optimize the admixture proportions, we used the Nelder-Mead package R package (https://github.com/mailund/admixture_graph) for inferring (https://cran.r-project.org/web/packages/neldermead/index.html). graphs and graph parameters from vectors of observed statistics and Exploring the space of admixture graph topologies applied this approach to the baboon data. We explored the space of graph topologies with a brute-force approach, We used qpDstats from the ADMIXTOOLS package to compute bounding the number of allowed admixture events. Because of the large estimates of f with corresponding Z values for all quartets (W, X, Y, number of possible graphs, we could not exhaustively explore all the Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 8 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE topologies including all the species (six baboon species and the rhesus where qi ~ p(q|T), but this estimator has a large variance since most macaque outgroup). Instead, we used various heuristics. parameters drawn from the prior distribution have a very low likeli- We could explore all the trees with seven leaves. For admixture hood. Instead, we used the samples from the posterior distribution to graphs with a single admixture event, we could only explore all the estimate the likelihood using an importance sampler. graphs with six leaves; hence, we built all the graphs with one baboon Now, because ∫pðqjD;TÞ pðD;qjTÞ=pðDjTÞ pðqjTÞ pðDjq;TÞ dq = ∫ pðD;qjTÞ=pðqjTÞ dq = ∫pðDjTÞ dq ¼ pðDjTÞ 1 species missing and then reinserted the missing species into the 1 ∫pðqjTÞdq = pðDjTÞ ; we can estimate ½pðDjTÞ = ∫ pðDjq;TÞ pðqjD; TÞ 1 1 graph (such that inserting it cannot add a second admixture event). N For additional admixture events, we took a greedy approach and ex- dq ≈ N1 ∑ ½pðDjqi ; TÞ1 ; where qi ~ p(q|D, T). plored all the graphs reachable by adding a new admixture event to i¼1 the best graphs with one less event (although we have no guarantee Identification of admixture through asymmetric that the optimal graph with n events is necessarily an extension in this allele sharing way of the optimal graph with n - 1 events). To further investigate possible admixture among baboon species, we Estimating graph parameters and testing models identified asymmetries in informative site patterns within the 15 ba- We developed a Metropolis-Hastings MCMC (Markov Chain Monte boon diversity samples. This approach detects tracts of recent introgres- Carlo) to sample the posterior of graph parameters given observed f- sion between the different baboon lineages. Nucleotide sites where a statistics. By sampling the posterior of graph parameters, we obtained derived variant is shared by two species to the exclusion of another estimates of admixture proportions and the uncertainty in these esti- are informative of the underlying gene tree. The two informative site mates. In addition, we can use the posterior samples in an importance patterns, which group together species that are not the most closely re- Downloaded from http://advances.sciencemag.org/ on February 10, 2019 sampler to obtain the likelihood of the observed statistics given a graph lated species, arise from either recurrent mutations or ILS. These pat- topology by integrating over the parameters of the graph. terns are expected to show equal frequencies if the mutation rate is Sampling graph parameters constant across the tree and if there has been no asymmetric admixture Let D denote the observed data (the estimated f-statistics), T a given between a non-sister species and the two sister species tested. A strong graph topology, and q the graph parameters of graph T (edge lengths and consistent asymmetry in the frequency of the two site patterns and admixture proportions). The purpose of the MCMC is to sample supporting alternative gene trees may thus result from asymmetric ad- over the posterior distribution of graph parameters given the observed mixture. To identify tracts of recent admixture, we counted informative data and the graph topology, i.e., sample from p(q|D, T). Since we can site patterns in 1-Mb windows along the alignments of individual gen- compute the likelihood of a parameter point, p(D|q, T), up to a normal- omes from all species trios of baboons. We computed the 0.99 quantile ization factor, we can use the Metropolis-Hastings algorithm to construct of counts of the two alternative site patterns across 1-Mb windows (the an MCMC that samples over the posterior distribution. We trans- two site patterns that support gene trees with a topology different from formed the parameter space to make the proposal distribution symmet- the species tree). We then call admixture tracts as consecutive 1-Mb bins ric (thus, the acceptance probability for the Metropolis-Hastings step is where the count of one alternative site pattern was above the 0.99 quan- simply the posterior ratio) and to ensure that the parameters are all legal tile of the other alternative site pattern. for their interpretation in the graph framework. Edge lengths, which must be positive, were log-transformed, and admixture proportions, Polymorphic Alu detection and characterization which must fall within the unit interval, were transformed with the A computational analysis was performed to identify full-length lineage inverse normal cumulative distribution function. After transformation, specific AluY sequences in the olive baboon (Panu_3.0), rhesus ma- we used an adaptive algorithm to construct the proposal distribution caque (Mmul_8.0.1), African green monkey (chlSab2), orangutan based on correlations in the previously sampled variables. For each (P_pygmaeus 2.0.2), chimpanzee (Pan_tro3), and human (GRCH38/ graph, we ran three independent chains with 10,000 steps and we hg38) reference genomes, as previously described by Steely et al. (53). checked convergence by comparing the distributions from the inde- Initial screen for polymorphic Alu insertions pendent chains. Since edge lengths are measured in terms of drift, they Two methods were used to identify 494 informative Alu insertions, a do not have a simple interpretation as time parameters. We therefore subset of young elements currently polymorphic within the genus considered them as nuisance parameters in the MCMC analysis. We Papio. The first method used BLAT to align AluY sequences obtained used them to test convergence of the Markov chains, but in the results from the P. anubis assembly (Panu_2.0) against M. mulatta (RheMac2) presented here, we focused on estimates of admixture proportions. and H. sapiens (GRCh38/hg38). Insertions present in P. anubis yet ab- Calculating graph topology likelihoods sent from the other two assemblies were subjected to polymerase chain To compare two graph topologies, we can compute the topology like- reaction (PCR) analysis to determine whether they were polymorphic lihood p(D|T) for a given topology T. Using this likelihood estimate, we across baboon species. A total of 187 loci were retained after this analysis. 1Þ can compare two topologies using the Bayes factor K T 1 ;T 2 ¼ pðDjTpðDjT 2 Þ , Computational analysis of diversity samples which captures the relative support the data provides for one topology over In our second method, whole-genome sequence data generated from another: If, e.g., K T 1 ;T 2 ¼ 10, we would consider T1 the more likely six of the diversity panel baboons (16098, 28547, 28755, 34472, 34474, topology unless a priori topology T2 was at least 10 times more likely and 97124) were aligned to P. anubis genome (Panu_2.0), as previously than T1. described by Jordan et al. (54). Given data D and topology T, the likelihood of T is computed by Primer design integrating over all the graph parameters for the topology p(D|T) = Oligonucleotide primers for locus-specific PCR were designed as re- ∫p(D, q|T) dq = ∫p(D|q, T) p(q|T) dq. This integral can be approximated ported in (53). by sampling from the prior distribution of graph parameters and com- Sanger sequencing N puting the mean likelihood, ∫pðDjq; TÞpðqjTÞdq ≈ N1 ∑ pðDjqi ; TÞ, Following PCR analysis of candidate Alu insertion polymorphisms on i¼1 the DNA panel of baboons, a small number of loci required Sanger Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 9 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE sequencing for clarification. The first category included eight loci For each locus, we extracted the sequence for all baboon diversity ascertained from either a yellow or kinda baboon genome, computa- panel individuals, replacing heterozygous sites with their corresponding tionally absent from the reference genome Panu_2.0, but for which IUPAC (International Union of Pure and Applied Chemistry) codes. PCR results indicated a filled site, or Alu present amplicon size, for re- To get an outgroup sequence for each locus, we compared the baboon ference DNA sample 27861. These were classified as possible “false- reference genome (Panu_2.0) to the rhesus macaque reference genome negative” events (supposed to be absent but were not), and PCR fragments sequence (rheMac2) using megablast, retaining hits with e values less were sequenced from 27861 and the ascertained individual to confirm than 1 × 10-100 and at least 95% identity. Only loci with a single rhesus the existence of a shared insertion event. The second category included match were retained. All analyses used custom scripts and BEDtools, nine loci in which the Alu present PCR product was either larger or SAMtools, and VCFtools. We aligned the sequences at each locus with smaller than the predicted filled size amplicon, or displayed both size Muscle using default parameters. bands, in one or more Papio species, but not all. All applicable PCR frag- We inferred a gene tree for each locus using MrBayes 3.2.1 (50), ments were sequenced to confirm that the ascertained Alu element of setting the outgroup to the rhesus sequence. For each alignment, we interest was present and to determine what the extra sequence con- used a GTR+G model of molecular evolution, and we ran MrBayes tained. Four PCR fragments per locus were gel-purified using a Wizard twice for 1,000,000 generations, sampling every 100. We assessed SV gel purification kit (Promega Corporation, Madison, WI, USA, cat- convergence by checking statistics in MrBayeslog (LnL, PSRF, and alog A9282) according to the manufacturer’s instruction. Cycle sequenc- average SD of split frequencies <0.01) and by using Tracer v.1.5 to ing was performed, and resulting products were cleaned by standard estimate the effective sample size as >200 and to compare the ethanol precipitation. Capillary electrophoresis was performed on an performance of the independent analyses. After checking for conver- Downloaded from http://advances.sciencemag.org/ on February 10, 2019 ABI 3130xl Genetic Analyzer (Applied Biosystems Inc., Foster City, gence, we summarized the posterior distribution of trees after re- CA) and evaluated using ABI software Sequence Scanner v1.0. Sequence moving the first 25% of generations. alignment figures were constructed in BioEdit, and a consensus sequence Filtration for phylogenetic information content and BCA for each locus was determined from the multiple forward and reverse We used the program mbsum from BUCKy v 1.4.2 to summarize the Sanger sequences obtained for each locus (53). posterior tree output from MrBayes for each locus (31, 55). As is Phylogenetic analysis expected with a recent radiation, many loci had limited phylogenetic A phylogenetic tree was created using the larger dataset of polymorphic information content, decreasing the signal-to-noise ratio and increasing Alu elements as characters. If an Alu insertion was fixed present (homo- concordance analysis runtime. To quickly remove loci with limited signal, zygous) at a particular locus, it was coded as “1”. If the insertion was we filtered loci based on the frequency of the highest supported tree in the fixed absent at a particular locus, it was coded as “0”. Insertions that MrBayes tally of tree topologies. A locus with no information is expected to were found to be heterozygous in an individual were coded as “1,0”. output a flat distribution of random topologies, each appearing once. In con- If a locus could not be resolved through PCR, then it was coded as a trast, a locus with strong signal will have the same topology occur multiple “?”. Mesquite 3.04 (53, 54) was used to create a Dollo parsimony matrix times. We removed loci if the most frequently supported topology with all characters set to the character type Dollo.up. A heuristic search occurred in fewer than 10% of trials. We then ran the main BUCKy was completed using PAUP* 4.0a147 (54) with a total of 10,000 boot- program with default parameters to infer CF summary statistics that strap replicates. The majority-rule tree does not include bootstrap val- describe the proportion of trees that contain a particular clade. ues below 50% for any branches, but does include values for consistency Genic tree analysis of sequences containing annotated index, retention index, and homoplasy index. A neighbor-joining phy- protein-coding genes logenetic tree was also created using the same dataset to illustrate the We also analyzed 3267 chromosomal segments that each contain one overall topology. The trees were produced in PAUP* and visualized annotated protein-coding gene. Segments were selected based on re- using FigTree 1.4.2 (http://tree.bio.ed.ac.uk/software/figtree/). fGene tables within the UCSC Genome Browser. The inference of local tree topologies was performed as for the loci discussed above that do not Reconstructing phylogeny and admixture from gene contain protein-coding genes. We began with 3267 genic segments, but tree analyses following filtering for phylogenetic signal, length, and other criteria, we Locus-specific phylogeny from loci not containing obtained final results for 2201. For each of these genic segments, we com- protein-coding genes puted pairwise Euclidean distances using the Kendall and Colijn metric To infer baboon phylogeny from a series of independent putatively (56) and then performed PCA (Principal Component Analysis) on this neutral loci (gene trees), we identified and excluded bases with any distance matrix to group trees into six clusters based on tree similarity. of the following characteristics: genic regions, based on refGene anno- We chose to further investigate the first three clusters. We performed a tations accessed via the refGene table of UCSC’s Genome Browser; GO overrepresentation test using as a reference list the genes for all trees bases with coverage <7× for any individual in the diversity panel; bases that survived filtering. Results of this for GO biological processes and mo- with any missing genotype; bases with read depth above the 95th lecular functions are reported in table S10. percentile; CpG sites; bases within 3 bp of an indel; repetitive DNA as designated by RepeatMasker (using the open-3-3-0 version of CoalHMM trees RepeatMasker with sensitive setting RepBase library release 20110920) CoalHMMs (32, 33, 57) exploit the Markov approximation to the se- and Tandem Repeats Finder (period of 12 or less); and bases within quential coalescent process (58). These models can be constructed to 100 bp of a phastCons element. For the bases that passed filtration, we capture various demographic scenarios in the ancestry of a set of concatenated nearby loci separated by gaps of less than 1 kb using sampled chromosomes. We have constructed a new model to infer BEDtools. We then retained only loci of size 1 to 100 kb for further parameters relevant for chromosomes from an admixed population analyses to maximize information content while still reducing the and one or two populations related to the donor populations of the ad- chance of unappreciated recombination. mixture event. Denote the populations in the model A, B, and C, where Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 10 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE C is descended from the admixed population and A and B are related to Section S12. CoalHMMs of admixture trees and events Section S13. Locus-specific phylogenetic trees for chromosomal segments containing the two donor populations. Parameters of the model include divergence annotated genes times between all pairs of populations and the admixture proportions Fig. S1. Panu3.0 genome assembly process. for the admixed population. Fig. S2. Workflow used in variant calling pipeline. Kinda admixture Fig. S3. Details regarding SNV calls. We first consider the most probable graph with a single admixture Fig. S4. Maximum likelihood and Bayesian phylogenetic trees based on SNV data. Fig. S5. Test of phylogeny reconstruction using PoMo. event. As an example, we consider P. kindae as the admixed population, Fig. S6. Identification of admixture using f-statistics. C. We have two choices for population A, P. cynocephalus and P. ursinus, Fig. S7. Evidence for admixture from haplotyping sharing. and two choices for population B, P. hamadryas and P. papio. With these Fig. S8. A cladogram of Papio individuals from the diversity panel. choices, since each of the populations considered for A and B has two Fig. S9. Bayesian concordance analysis. Fig. S10. Bootstrap analysis of timing of divergence events. samples, we could choose independent chromosomes to compare for re- Fig. S11. Confidence intervals for baboon admixture proportions. plication of results. With these species, we can estimate the time param- Fig. S12. Results for simulated admixture analysis. eters for specific evolutionary events: (i) the north-south clade split, Fig. S13. Results for correction factor adjustment of admixture history. (ii) the split between hamadryas/papio and the species that hybridized Fig. S14. Model used to estimate specific divergence and admixture history. to become P. kindae, (iii) the cynocephalus/ursinus split, (iv) the split be- Fig. S15. Unbiased estimates dating divergences and admixture events. Fig. S16. Phylogeny representing cluster 1 genic regions. tween ursinus and the other species that hybridized to become P. kindae, Fig. S17. Phylogeny representing cluster 2 genic regions. and (v) the time of the admixture that formed P. kindae. Not all of these Fig. S18. Phylogeny representing cluster 3 genic regions. time points can be estimated by all triplets of species, but all can be es- Table S1. Assembly statistics. Downloaded from http://advances.sciencemag.org/ on February 10, 2019 timated from at least two sets of species triplets, giving us four Table S2. Annotation of baboon genome assemblies. Table S3. DNA samples used for diversity analysis. independent estimates. Table S4. SNV variation among 15 Papio baboons and a gelada. Simulation test of goodness of fit and debiasing estimates Table S5. Full-length AluY insertions and lineage-specific insertions in primate genomes. To examine which parameters are likely to be biased, and by how much, Table S6. Effect of admixture on branch lengths measured in substitutions per site. we simulated data with parameters in a grid of time points around the Table S7. Bayes factors comparing alternate phylogenies. estimated points and estimated the parameters from these simulated Table S8. Divergence time estimates across triplets. Table S9. Admixture proportion estimates across triplets. data. Figure S12 shows the results with the estimated time points and Table S10. GO terms associated with genes falling in clusters 1 to 3 of genic regions. the admixture proportions together with simulated data, where the sim- References (61–78) ulated values are shown as black points and the corresponding esti- mated parameters are shown as red error bars (these error bars are wider since we used smaller datasets for the simulated data for compu- REFERENCES AND NOTES tational reasons). 1. O. Seehausen, R. K. Butlin, I. Keller, C. E. Wagner, J. W. Boughman, P. A. Hohenlohe, C. L. Peichel, G.-P. Saetre, C. Bank, Å. Brännström, A. Brelsford, C. S. Clarkson, Dating divergence and admixture events within F. Eroukhmanoff, J. L. Feder, M. C. Fischer, A. D. Foote, P. Franchini, C. D. Jiggins, the CoalHMM F. C. Jones, A. K. Lindholm, K. Lucek, M. E. Maan, D. A. Marques, S. H. Martin, B. Matthews, J. I. Meier, M. Möst, M. W. Nachman, E. Nonaka, D. J. Rennison, J. Schwarzer, E. T. Watson, The CoalHMM produces a phylogenetic tree with relative dates for A. M. Westram, A. Widmer, Genomics and the origin of species. Nat. Rev. Genet. 15, nodes within the tree scaled to nucleotide substitutions. We first as- 176–192 (2014). sumed that the initial divergence of northern and southern clades of 2. J. B. W. Wolf, H. Ellegren, Making sense of genomic islands of differentiation in light of baboons occurred 2.0 Ma ago, as suggested by analyses of mtDNA di- speciation. Nat. Rev. Genet. 18, 87–100 (2017). vergences (12). The ratios of inferred lineage-specific nucleotide substi- 3. D. Otte, J. A. Endler, Speciation and Its Consequences (Sinauer Associates, Inc., 1989). 4. E. Mayr, Systematics and the Origin of Species (Columbia Univ. Press, 1942). tutions were then used to calculate absolute dates for nodes in the 5. M. L. Arnold, Divergence with Genetic Exchange (Oxford Univ. Press, 2015). CoalHMM tree. As an alternative approach to dating, we used the es- 6. J. Mallet, N. Besansky, M. W. Hahn, How reticulated are species? Bioessays 38, 140–149 timated baboon mutation rate of 0.9 × 10-8 per base pair per generation (2016). (see section S2). With a generation time of 11 years, the absolute date for 7. L. A. Cox, A. G. Comuzzie, L. M. Havill, G. M. Karere, K. D. Spradling, M. C. Mahaney, P. W. Nathanielsz, D. P. Nicolella, R. E. Shade, S. Voruganti, J. L. VandeBerg, Baboons as a each node within the CoalHMM tree was calculated using the branch model to study genetics and epigenetics of human disease. ILAR J. 54, 106–121 (2013). lengths determined through coalescent modeling, assuming the indi- 8. C. M. Kammerer, L. A. Cox, M. C. Mahaney, J. Rogers, R. E. Shade, Sodium-lithium cated mutation rate and generation time. countertransport activity is linked to chromosome 5 in baboons. Hypertension 37, 398–402 (2001). 9. C. J. Jolly, A proper study for mankind: Analogies from the Papionin monkeys and their implications for human evolution. Am. J. Phys. Anthropol. 116 (suppl. 33), 177–204 (2001). SUPPLEMENTARY MATERIALS 10. C. J. Jolly, in Species, Species Concepts and Primate Evolution, W. H. Kimbel, L. B. Martin, Supplementary material for this article is available at http://advances.sciencemag.org/cgi/ Eds. (Plenum Press, 1993), pp. 67–107. content/full/5/1/eaau6947/DC1 11. L. Swedell, in Primates in Perspective, C. J. Campbell, A. Fuentes, K. C. MacKinnon, Section S1. Rationale for baboon taxonomy and nomenclature S. K. Bearder, R. M. Stumpf, Eds. (Oxford Univ. Press, ed. 2, 2011). Section S2. Rationale for mutation rate used in PSMC analyses 12. D. Zinner, J. Wertheimer, R. Liedigk, L. F. Groeneveld, C. Roos, Baboon phylogeny as inferred Section S3. Sequencing and assembly of olive baboon genome from complete mitochondrial genomes. Am. J. Phys. Anthropol. 150, 133–140 (2013). Section S4. Annotation and gene content of the baboon genome 13. S. C. Antón, R. Potts, L. C. Aiello, Human evolution. Evolution of early Homo: An integrated Section S5. Identification of SNVs and small indels within baboon species biological perspective. Science 345, 1236828 (2014). Section S6. Validation of species identity within the diversity panel 14. D. Zinner, U. Buba, S. Nash, C. Roos, in Primates of Gashaka, V. Sommer, C. Roos, Eds. Section S7. Lineage-specific Alu insertion in OWMs and hominoids (Springer, 2011), pp. 267–306. Section S8. Alternative methods for constructing phylogenetic trees 15. D. Zinner, C. Keller, J. W. Nyahongo, T. M. Butynski, Y. A. de Jong, L. Pozzi, S. Knauf, Section S9. Identification of admixture through asymmetric allele sharing R. Liedigk, C. Roos, Distribution of mitochondrial clades and morphotypes of baboons Section S10. Polymorphic AluY insertions across Papio species Papio spp. (Primates: Cercopithecidae) in eastern Africa. J. East African Nat. Hist. 104, Section S11. Bayesian concordance analyses of gene trees devoid of coding sequences 143–168 (2015). Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 11 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE 16. J. Fischer, G. H. Kopp, F. Dal Pesco, A. Goffe, K. Hammerschmidt, U. Kalbitzer, M. Klapproth, 40. H. Kummer, in Primates: Studies in Adaptation and Variability, P. C. Jay, Ed. (Holt, Rinehart & P. Maciej, I. Ndao, A. Patzelt, D. Zinner, Charting the neglected West: The social system Winston, 1968), pp. 293–312. of Guinea baboons. Am. J. Phys. Anthropol. 162 (suppl. 63), 15–31 (2017). 41. C. M. Bergey, J. E. Phillips-Conroy, T. R. Disotell, C. J. Jolly, Dopamine pathway is highly 17. M. J. E. Charpentier, M. C. Fontaine, E. Cherel, J. P. Renoult, T. Jenkins, L. Benoit, N. Barthès, diverged in primate species that differ markedly in social behavior. Proc. Natl. Acad. Sci. U.S.A. S. C. Alberts, J. Tung, Genetic structure in a dynamic baboon hybrid zone corroborates 113, 6178–6181 (2016). behavioural observations in a hybrid population. Mol. Ecol. 21, 715–731 (2012). 42. C. J. Jolly, J. E. Phillips-Conroy, J. R. Kaplan, J. J. Mann, Cerebrospinal fluid monoaminergic 18. C. J. Jolly, A. S. Burrell, J. E. Phillips-Conroy, C. Bergey, J. Rogers, Kinda baboons (Papio metabolites in wild Papio anubis and P. hamadryas are concordant with taxon-specific kindae) and grayfoot chacma baboons (P. ursinus griseipes) hybridize in the Kafue river behavioral ontogeny. Int. J. Primatol. 29, 1549–1566 (2008). valley, Zambia. Am. J. Primatol. 73, 291–303 (2011). 43. C. M. Kammerer, M. L. Sparks, J. Rogers, Effects of age, sex, and heredity on measures of 19. T. J. Bergman, J. E. Phillips-Conroy, C. J. Jolly, Behavioral variation and reproductive bone mass in baboons (Papio hamadryas). J. Med. Primatol. 24, 236–242 (1995). success of male baboons (Papio anubis × Papio hamadryas) in a hybrid social group. 44. P. Kochunov, D. C. Glahn, P. T. Fox, J. L. Lancaster, K. Saleem, W. Shelledy, K. Zilles, Am. J. Primatol. 70, 136–147 (2008). P. M. Thompson, O. Coulon, J. F. Mangin, J. Blangero, J. Rogers, Genetics of primary 20. S. Pääbo, The diverse origins of the human gene pool. Nat. Rev. Genet. 16, 313–314 cerebral gyrification: Heritability of length, depth and area of primary sulci in an (2015). extended pedigree of Papio baboons. Neuroimage 53, 1126–1134 (2010). 21. B. Vernot, S. Tucci, J. Kelso, J. G. Schraiber, A. B. Wolf, R. M. Gittelman, M. Dannemann, 45. C. M. Moore, C. Janish, C. A. Eddy, G. B. Hubbard, M. M. Leland, J. Rogers, Cytogenetic and S. Grote, R. C. McCoy, H. Norton, L. B. Scheinfeldt, D. A. Merriwether, G. Koki, J. S. Friedlaender, fertility studies of a rheboon, rhesus macaque (Macaca mulatta) × baboon (Papio J. Wakefield, S. Pääbo, J. M. Akey, Excavating Neandertal and Denisovan DNA from the hamadryas) cross: Further support for a single karyotype nomenclature. Am. J. Phys. genomes of Melanesian individuals. Science 352, 235–239 (2016). Anthropol. 110, 119–127 (1999). 22. W. C. Warren, A. J. Jasinska, R. García-Pérez, H. Svardal, C. Tomlinson, M. Rocchi, N. Archidiacono, 46. A. C. English, S. Richards, Y. Han, M. Wang, V. Vee, J. Qu, X. Qin, D. M. Muzny, J. G. Reid, O. Capozzi, P. Minx, M. J. Montague, K. Kyung, L. D. W. Hillier, M. Kremitzki, T. Graves, C. Chiang, K. C. Worley, R. A. Gibbs, Mind the gap: Upgrading genomes with Pacific Biosciences J. Hughes, N. Tran, Y. Huang, V. Ramensky, O.-W. Choi, Y. J. Jung, C. A. Schmitt, N. Juretic, RS long-read sequencing technology. PLOS ONE 7, e47768 (2012). J. Wasserscheid, T. R. Turner, R. W. Wiseman, J. J. Tuscher, J. A. Karl, J. E. Schmitz, R. Zahn, 47. B. J. Walker, T. Abeel, T. Shea, M. Priest, A. Abouelliel, S. Sakthikumar, C. A. Cuomo, Downloaded from http://advances.sciencemag.org/ on February 10, 2019 D. H. O’Connor, E. Redmond, A. Nisbett, B. Jacquelin, M. C. Müller-Trutwin, J. M. Brenchley, Q. Zeng, J. Wortman, S. K. Young, A. M. Earl, Pilon: An integrated tool for comprehensive M. Dione, M. Antonio, G. P. Schroth, J. R. Kaplan, M. J. Jorgensen, G. W. C. Thomas, microbial variant detection and genome assembly improvement. PLOS ONE 9, e112963 M. W. Hahn, B. J. Raney, B. Aken, R. Nag, J. Schmitz, G. Churakov, A. Noll, R. Stanyon, (2014). D. Webb, F. Thibaud-Nissen, M. Nordborg, T. Marques-Bonet, K. Dewar, G. M. Weinstock, 48. X. Liu, S. White, B. Peng, A. D. Johnson, J. A. Brody, A. H. Li, Z. Huang, A. Carroll, P. Wei, R. K. Wilson, N. B. Freimer, The genome of the vervet (Chlorocebus aethiops sabaeus). R. Gibbs, R. J. Klein, E. Boerwinkle, WGSA: An annotation pipeline for human genome Genome Res. 25, 1921–1933 (2015). sequencing studies. J. Med. Genet. 53, 111–112 (2016). 23. N. De Maio, D. Schrempf, C. Kosiol, PoMo: An allele frequency-based approach for species 49. L.-T. Nguyen, H. A. Schmidt, A. von Haeseler, B. Q. Minh, IQ-TREE: A fast and effective tree estimation. Syst. Biol. 64, 1018–1031 (2015). stochastic algorithm for estimating maximum-likelihood phylogenies. Mol. Biol. Evol. 32, 24. D. Schrempf, B. Q. Minh, N. De Maio, A. von Haeseler, C. Kosiol, Reversible polymorphism- 268–274 (2015). aware phylogenetic models and their application to tree inference. J. Theor. Biol. 407, 50. F. Ronquist, M. Teslenko, P. van der Mark, D. L. Ayres, A. Darling, S. Höhna, B. Larget, L. Liu, 362–370 (2016). M. A. Suchard, J. P. Huelsenbeck, MrBayes 3.2: Efficient Bayesian phylogenetic inference 25. S. C. Alberts, J. Altmann, Immigration and hybridization patterns of yellow and anubis and model choice across a large model space. Syst. Biol. 61, 539–542 (2012). baboons in and around Amboseli, Kenya. Am. J. Primatol. 53, 139–154 (2001). 51. G. Ewing, J. Hermisson, MSMS: A coalescent simulation program including 26. J. Tung, M. J. E. Charpentier, D. A. Garfield, J. Altmann, S. C. Alberts, Genetic evidence reveals recombination, demographic structure and selection at a single locus. Bioinformatics 26, temporal change in hybridization patterns in a wild baboon population. Mol. Ecol. 17, 2064–2065 (2010). 1998–2011 (2008). 52. N. Patterson, P. Moorjani, Y. Luo, S. Mallick, N. Rohland, Y. Zhan, T. Genschoreck, 27. J. D. Wall, S. A. Schlebusch, S. C. Alberts, L. A. Cox, N. Snyder-Mackler, K. A. Nevonen, T. Webster, D. Reich, Ancient admixture in human history. Genetics 192, 1065–1093 L. Carbone, J. Tung, Genomewide ancestry and divergence patterns from low-coverage (2012). sequencing data reveal a complex history of admixture in wild baboons. Mol. Ecol. 25, 53. C. J. Steely, J. N. Baker, J. A. Walker, C. D. Loupe III, Analysis of lineage-specific Alu 3469–3483 (2016). subfamilies in the genome of the olive baboon, Papio anubis. Mob DNA 9, 10 (2018). 28. J. E. Phillips-Conroy, C. J. Jolly, F. L. Brett, Characteristics of hamadryas-like male baboons 54. V. E. Jordan, J. A. Walker, T. O. Beckstrom, C. J. Steely, C. L. McDaniel, C. P. St. Romain; living in anubis baboon troops in the Awash hybrid zone, Ethiopia. Am. J. Phys. Anthropol. 86, Baboon Genome Analysis Consortium, K. C. Worley, J. Phillips-Conroy, C. J. Jolly, J. Rogers, 353–368 (1991). M. K. Konkel, M. A. Batzer, A computational reconstruction of Papio phylogeny using 29. D. A. Ray, J. Xing, A.-H. Salem, M. A. Batzer, SINEs of a nearly perfect character. Syst. Biol. Alu insertion polymorphisms. Mob DNA 9, 13 (2018). 55, 928–935 (2006). 55. C. Ané, B. Larget, D. A. Baum, S. D. Smith, A. Rokas, Bayesian estimation of concordance 30. F. K. Mendes, M. W. Hahn, Gene tree discordance causes apparent substitution rate among gene trees. Mol. Biol. Evol. 24, 412–426 (2007). variation. Syst. Biol. 65, 711–721 (2016). 56. M. Kendall, C. Colijn, Mapping phylogenetic trees to reveal distinct patterns of evolution. 31. B. R. Larget, S. K. Kotha, C. N. Dewey, C. Ané, BUCKy: Gene tree/species tree reconciliation Mol. Biol. Evol. 33, 2735–2743 (2016). with Bayesian concordance analysis. Bioinformatics 26, 2910–2911 (2010). 57. T. Mailund, A. E. Halager, M. Westergaard, J. Y. Dutheil, K. Munch, L. N. Andersen, G. Lunter, 32. J. Y. Dutheil, G. Ganapathy, A. Hobolth, T. Mailund, M. K. Uyenoyama, M. H. Schierup, K. Prüfer, A. Scally, A. Hobolth, M. H. Schierup, A new isolation with migration model Ancestral population genomics: The coalescent hidden Markov model approach. Genetics along complete genomes infers very different divergence processes among closely 183, 259–274 (2009). related great ape species. PLOS Genet. 8, e1003125 (2012). 33. T. Mailund, J. Y. Dutheil, A. Hobolth, G. Lunter, M. H. Schierup, Estimating divergence 58. G. A. T. McVean, N. J. Cardin, Approximating the coalescent with recombination. Philos. time and ancestral effective population size of Bornean and Sumatran orangutan Trans. R. Soc. Lond. B Biol. Sci. 360, 1387–1393 (2005). subspecies using a coalescent hidden Markov model. PLOS Genet. 7, e1001319 (2011). 59. E. Delson, C. J. Terranova, W. L. Jungers, E. J. Sargis, N. G. Jablonski, P. C. Dechow, 34. J. Tung, E. A. Archie, J. Altmann, S. C. Alberts, Cumulative early life adversity predicts Body Mass in Cercopithecidae (Primates, Mammalia): Estimation and Scaling in Extinct and longevity in wild baboons. Nat. Commun. 7, 11181 (2016). Extant Taxa (Anthropological Papers of the American Museum of Natural History, 2000). 35. H. Li, R. Durbin, Inference of human population history from individual whole-genome 60. D. P. Locke, L. W. Hillier, W. C. Warren, K. C. Worley, L. V. Nazareth, D. M. Muzny, S.-P. Yang, sequences. Nature 475, 493–496 (2011). Z. Wang, A. T. Chinwalla, P. Minx, M. Mitreva, L. Cook, K. D. Delehaunty, C. Fronick, H. Schmidt, 36. N. G. Jablonski, S. Frost, in Cenozoic Mammals of Africa, L. Werdelin, W. J. Sanders, Eds. L. A. Fulton, R. S. Fulton, J. O. Nelson, V. Magrini, C. Pohl, T. A. Graves, C. Markovic, A. Cree, (University of California Press, 2010), pp. 393–428. H. H. Dinh, J. Hume, C. L. Kovar, G. R. Fowler, G. Lunter, S. Meader, A. Heger, C. P. Ponting, 37. S. Dennenmoser, F. J. Sedlazeck, E. Iwaszkiewicz, X.-Y. Li, J. Altmüller, A. W. Nolte, Copy T. Marques-Bonet, C. Alkan, L. Chen, Z. Cheng, J. M. Kidd, E. E. Eichler, S. White, S. Searle, number increases of transposable elements and protein-coding genes in an invasive A. J. Vilella, Y. Chen, P. Flicek, J. Ma, B. Raney, B. Suh, R. Burhans, J. Herrero, D. Haussler, R. Faria, fish of hybrid origin. Mol. Ecol. 26, 4712–4724 (2017). O. Fernando, F. Darré, D. Farré, E. Gazave, M. Oliva, A. Navarro, R. Roberto, O. Capozzi, 38. R. J. W. O’Neill, M. J. O’Neill, J. A. M. Graves, Undermethylation associated with N. Archidiacono, G. Della Valle, S. Purgato, M. Rocchi, M. K. Konkel, J. A. Walker, B. Ullmer, retroelement activation and chromosome remodelling in an interspecific mammalian M. A. Batzer, A. F. A. Smit, R. Hubley, C. Casola, D. R. Schrider, M. W. Hahn, V. Quesada, hybrid. Nature 393, 68–72 (1998). X. S. Puente, G. R. Ordoñez, C. López-Otín, T. Vinar, B. Brejova, A. Ratan, R. S. Harris, W. Miller, 39. R. R. Ackermann, L. Schroeder, J. Rogers, J. M. Cheverud, Further evidence for phenotypic C. Kosiol, H. A. Lawson, V. Taliwal, A. L. Martins, A. Siepel, A. RoyChoudhury, X. Ma, signatures of hybridization in descendant baboon populations. J. Hum. Evol. 76, 54–62 J. Degenhardt, C. D. Bustamante, R. N. Gutenkunst, T. Mailund, J. Y. Dutheil, A. Hobolth, (2014). M. H. Schierup, O. A. Ryder, Y. Yoshinaga, P. J. de Jong, G. M. Weinstock, J. Rogers, E. R. Mardis, Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 12 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE R. A. Gibbs, R. K. Wilson, Comparative and demographic analysis of orang-utan genomes. M. M. Munidasa, L. N. Nazareth, I. N. Newsham, D. N. Ngo, L. N. Nguyen, P. Nguyen, T. N. Nguyen, Nature 469, 529–533 (2011). N. N. Nguyen, M. Nwaokelemeh, M. O. Obregon, G. O. Okwuonu, F. O. Ongeri, C. O. Onwere, 61. J. Cracraft, in Speciation and Its Consequences, D. Otte, J. A. Endler, Eds. (Sinauer I. O. Osifeso, A. P. Parra, S. P. Patil, A. P. Perez, Y. P. Perez, C. P. Pham, E. Primus, L.-L. Pu, M. P. Puazo, Associates, Inc., 1989). J. Q. Quiroz, S. Richards, J. R. Rouhana, M. R. Ruiz, S.-J. Ruiz, N. S. Saada, J. S. Santibanez, 62. J. A. Coyne, H. A. Orr, The evolutionary genetics of speciation. Philos. Trans. R. Soc. Lond. B M. S. Scheel, S. Scherer, B. S. Schneider, D. S. Simmons, I. S. Sisson, E. S. Skinner, N. Tabassum, Biol. Sci. 353, 287–305 (1998). L.-Y. Tang, A. Taylor, R. T. Thornton, J. T. Tisius, G. T. Toledanes, Z. T. Trejos, K. U. Usmani, 63. E. Mayr, Populations, Species and Evolution (Belknap Press of Harvard Univ. Press, 1963). R. V. Varghese, S. V. Vattathil, V. V. Vee, D. W. Walker, G. W. Weissenberger, C. W. White, 64. J. Cracraft, Species concepts and speciation analysis. Curr. Ornithol. 1, 159–187 (1983). K. Wilczek-Boney, A. W. Williams, K. Wilson, I. Woghiren, J. W. Woodworth, R. W. Wright, Y.-Q. Wu, 65. C. Groves, Primate Taxonomy (Smithsonian Institution Press, 2001). Y. Xin, Y. Zhang, Y. Z. Zhu, and X. Zou. The biomaterials for the DNA sequencing of the 66. P. Grubb, T. M. Butynski, J. F. Oates, S. K. Bearder, T. R. Disotell, C. P. Groves, reference P. anubis baboon and several of the diversity panel of baboons were provided by T. T. Struhsaker, Assessment of the diversity of African primates. Int. J. Primatol. 24, the Southwest National Primate Research Center, San Antonio, TX, which was supported 1301–1357 (2003). by a grant from the NIH Office of Research Infrastructure Programs (P51-OD011133). The 67. S. Boissinot, L. Alvarez, J. Giraldo-Ramirez, M. Tollis, Neutral nuclear variation in Baboons research reported here complied with governmental and IACUC regulations and guidelines. (genus Papio) provides insights into their evolutionary and demographic histories. Am. J. J.R. is also affiliated with the Wisconsin National Primate Research Center, Madison, WI. C.K. Phys. Anthropol. 155, 621–634 (2014). is also affiliated with the Institut für Populationsgenetik, Vetmeduni Vienna, Austria, and D.S. is 68. S. R. Frost, L. F. Marcus, F. L. Bookstein, D. P. Reddy, E. Delson, Cranial allometry, newly affiliated with Eötvös Lorand University Budapest and Max Perutz Laboratories Vienna. phylogeography, and systematics of large-bodied papionins (primates: Cercopithecinae) Funding: The sequencing and analysis activities at the Human Genome Sequencing Center, inferred from geometric morphometric analysis of landmark data. Anat. Rec. A Discov. Baylor College of Medicine, were supported by NIH (NHGRI) grants U54-HG003273 and Mol. Cell. Evol. Biol. 275, 1048–1072 (2003). U54-HG006484 to R.A.G. and GAC grant 1 S10 RR026605 to J. G. Reid. This research was also 69. L. Ségurel, M. J. Wyman, M. Przeworski, Determinants of mutation rate variation in the supported by NIH grant R01-GM59290 to M.A.B.; grants from the Austrian Science Fund human germline. Annu. Rev. Genomics Hum. Genet. 15, 47–70 (2014). (FWF-P24551 and FWF-W1225) and Vienna Science and Technology Fund (WWTF-MA16-061) 70. N. Elango, J. Lee, Z. Peng, Y.-H. E. Loh, S. V. Yi, Evolutionary rate variation in Old World to C.K.; grants from the Wellcome Trust (WT108749/Z/15/Z) and EMBL to B.A., F.J.M., and M.M.; Downloaded from http://advances.sciencemag.org/ on February 10, 2019 monkeys. Biol. Lett. 5, 405–408 (2009). grants VEGA 1/0719/14 and APVV-14-0253 to T. Vinar (Consortium Member); MINECO/FEDER 71. T. K. Newman, C. J. Jolly, J. Rogers, Mitochondrial phylogeny and systematics of baboons grant, NIH U01-MH106874 grant, Howard Hughes International Early Career award, and Obra (Papio). Am. J. Phys. Anthropol. 124, 17–27 (2004). Social “La Caixa” award to T.M.-B.; NSF grants BNS83-03506 to J.P.-C.; NSF1029302 to J.P.-C., J.R., 72. N. A. O’Leary, M. W. Wright, J. R. Brister, S. Ciufo, D. Haddad, R. McVeigh, B. Rajput, and C.J.J.; BNS96-15150 to J.P.-C., C.J.J., and T.D.; and National Geographic Society and Leakey B. Robbertse, B. Smith-White, D. Ako-Adjei, A. Astashyn, A. Badretdin, Y. Bao, O. Blinkova, Foundation grants to J.P.-C. and C.J.J. E.E.E. is an investigator of the Howard Hughes Medical V. Brover, V. Chetvernin, J. Choi, E. Cox, O. Ermolaeva, C. M. Farrell, T. Goldfarb, T. Gupta, Institute. This work was supported, in part, by U.S. NIH grant HG002385 to E.E.E. Competing D. Haft, E. Hatcher, W. Hlavina, V. S. Joardar, V. K. Kodali, W. Li, D. Maglott, P. Masterson, interests: The authors declare that they have no competing interests. Data and materials K. M. McGarvey, M. R. Murphy, K. O’Neill, S. Pujar, S. H. Rangwala, D. Rausch, L. D. Riddick, availability: The raw read data, sample metadata, and other information for this genome assembly C. Schoch, A. Shkeda, S. S. Storz, H. Sun, F. Thibaud-Nissen, I. Tolstoy, R. E. Tully, A. R. Vatsan, project are available under Bioproject PRJNA260523 at www.ncbi.nlm.nih.gov. Additional C. Wallin, D. Webb, W. Wu, M. J. Landrum, A. Kimchi, T. Tatusova, M. DiCuccio, P. Kitts, information concerning RNA sequencing data is available at the Nonhuman Primate Reference T. D. Murphy, K. D. Pruitt, Reference sequence (RefSeq) database at NCBI: Current status, Transcriptome Project (http://nhprtr.org/). Further information regarding SNV and indel variation is taxonomic expansion, and functional annotation. Nucleic Acids Res. 44, D733–D745 (2016). available as a track on the UCSC browser (https://hgsc.bcm.edu/non-human-primates/baboon- 73. B. L. Aken, P. Achuthan, W. Akanni, M. R. Amode, F. Bernsdorff, J. Bhai, K. Billis, genome-project). Additional data related to this paper may be requested from the authors. D. Carvalho-Silva, C. Cummins, P. Clapham, L. Gil, C. García Girón, L. Gordon, T. Hourlier, S. E. Hunt, S. H. Janacek, T. Juettemann, S. Keenan, M. R. Laird, I. Lavidas, T. Maurel, Full membership of the Baboon Genome Analysis Consortium: Bronwen Aken1, W. McLaren, B. Moore, D. N. Murphy, R. Nag, V. Newman, M. Nuhn, C. K. Ong, A. Parker, Nicoletta Archidiacono2, Georgios Athanasiadis3, Mark A. Batzer4, Thomas O. Beckstrom4, M. Patricio, H. S. Riat, D. Sheppard, H. Sparrow, K. Taylor, A. Thormann, A. Vullo, B. Walts, Christina Bergey5,6, Konstantinos Billis1, Andrew Burrell5, Oronzo Capozzi2, Claudia R. Catacchio2, S. P. Wilder, A. Zadissa, M. Kostadima, F. J. Martin, M. Muffato, E. Perry, M. Ruffier, Jade Cheng3, Laura A. Cox7,8, Huyen H. Dinh9, Todd Disotell5, HarshaVardhan Doddapaneni9, D. M. Staines, S. J. Trevanion, F. Cunningham, A. Yates, D. R. Zerbino, P. Flicek, Ensembl Evan E. Eichler10,11, James Else12, Richard A. Gibbs9,13, Matthew W. Hahn14, Yi Han9, R. Alan Harris9,13, 2017. Nucleic Acids Res. 45, D635–D642 (2017). John Huddleston10, Shalini N. Jhangiani9, Clifford J. Jolly5, Vallmer E. Jordan4, Anis Karimpour-Fard15, 74. B. L. Aken, S. Ayling, D. Barrell, L. Clarke, V. Curwen, S. Fairley, J. Fernandez Banet, K. Billis, Miriam K. Konkel32, Gisela H. Kopp16,17, Viktoriya Korchina9, Carolin Kosiol18, Maximillian Kothe19, C. García Girón, T. Hourlier, K. Howe, A. Kähäri, F. Kokocinski, F. J. Martin, D. N. Murphy, Christie L. Kovar9, Lukas Kuderna20, Sandra L. Lee9, Kalle Leppälä3, Xiaoming Liu21, Yue Liu9, R. Nag, M. Ruffier, M. Schuster, Y. A. Tang, J.-H. Vogel, S. White, A. Zadissa, P. Flicek, Thomas Mailund3, Tomas Marques-Bonet20,22,23,33, Alessia Marra-Campanale2, Fergal J. Martin1, S. M. J. Searle, The Ensembl gene annotation system. Database 2016, baw093 (2016). Christopher E. Mason24, Marc de Manuel Montero20, Matthieu Muffato1, Kasper Munch3, 75. D. Zinner, L. F. Groeneveld, C. Keller, C. Roos, Mitochondrial phylogeography of baboons Shwetha Murali9, Donna M. Muzny9,13, Angela Noll19, Kymberleigh A. Pagel25, Antonio Palazzo2, (Papio spp.): Indication for introgressive hybridization? BMC Evol. Biol. 9, 83 (2009). Jera Pecotte7, Vikas Pejaver25, Jane Phillips-Conroy26, Lenore Pipes24, Veronica Searles Quick15, 76. D. F. Robinson, L. R. Foulds, Comparison of phylogenetics trees. Math. Biosci. 53, 131–147 Predrag Radivojac25, Archana Raja10, Brian J. Raney27, Muthuswamy Raveendran9, Karen Rice7, (1981). Mariano Rocchi2, Jeffrey Rogers9,13, Christian Roos19, Mikkel Heide Schierup3, Dominik Schrempf28, 77. M. K. Kuhner, J. Felsenstein, A simulation comparison of phylogeny algorithms under James M. Sikela15, Roscoe Stanyon29, Cody J. Steely4, Gregg W. C. Thomas14, Jenny Tung30, equal and unequal evolutionary rates. Mol. Biol. Evol. 11, 459–468 (1994). Mario Ventura2, Tauras P. Vilgalys30, Tomas Vinar31, Jerilyn A. Walker4, Lutz Walter19, Kim C. Worley9,13, 78. C. Xue, M. Raveendran, R. A. Harris, G. L. Fawcett, X. Liu, S. White, M. Dahdouli, D. R. Deiros, and Dietmar Zinner16. J. E. Below, W. Salerno, L. Cox, G. Fan, B. Ferguson, J. Horvath, Z. Johnson, S. Kanthaswamy, H. M. Kubisch, D. Liu, M. Platt, D. G. Smith, B. Sun, E. J. Vallender, F. Wang, R. W. Wiseman, 1 European Molecular Biology Laboratory, European Bioinformatics Institute, Hinxton, UK. R. Chen, D. M. Muzny, R. A. Gibbs, F. Yu, J. Rogers, The population genomics of rhesus 2 Department of Biology, University of Bari, Bari, Italy. 3Bioinformatics Research Centre, Aarhus macaques (Macaca mulatta) based on whole-genome sequences. Genome Res. 26, University, Aarhus, Denmark. 4Department of Biological Sciences, Louisiana State University, Baton 1651–1662 (2016). Rouge, LA, USA. 5Department of Anthropology, New York University, New York, NY, USA. 6 Department of Biological Sciences, Notre Dame University, South Bend, IN, USA. 7Southwest Acknowledgments: We acknowledge the contributions of the sequence production staff of National Primate Research Center, Texas Biomedical Research Institute, San Antonio, TX, USA. 8 the Human Genome Sequencing Center: K. A. Abraham, H. A. Akbar, S. A. Ali, U. A. Anosike, Department of Genetics, Texas Biomedical Research Institute, San Antonio, TX, USA. 9Human P. A. Aqrawi, F. A. Arias, T. A. Attaway, R. A. Awwad, C. B. Babu, D. B. Bandaranaike, P. B. Battles, Genome Sequencing Center, Baylor College of Medicine, Houston, TX, USA. 10Department of A. B. Bell, B. B. Beltran, D. B. Berhane-Mersha, C. B. Bess, C. B. Bickham, T. B. Bolden, K. Cardenas, Genome Sciences, University of Washington, Seattle, WA, USA. 11Howard Hughes Medical Institute, K. C. Carter, M. Cavazos, A. Chandrabose, S. Chao, D. C. Chau, A. C. Chavez, R. Chu, University of Washington, Seattle, WA, USA. 12Department of Pathology and Laboratory Medicine K. C. Clerc-Blankenburg, A. Cockrell, M. C. Coyle, A. Cree, M. D. Dao, M. L. Davila, L. D. Davy-Carroll, and Yerkes Primate Research Center, Emory University, Atlanta, GA, USA. 13Department of Molecular S. D. Denson, S. Dugan, V. Ebong, S. Elkadiri, S. F. Fernandez, P. F. Fernando, N. Flagg, L. F. Forbes, and Human Genetics, Baylor College of Medicine, Houston, TX, USA. 14Department of Biology, G. Fowler, C. F. Francis, L. F. Francisco, Q. F. Fu, R. Gabisi, R. G. Garcia, T. Garner, T. G. Garrett, Indiana University, Bloomington, IN, USA. 15Department of Biochemistry and Molecular Genetics, S. G. Gross, S. G. Gubbala, K. Hawkins, B. Hernandez, K. H. Hirani, M. H. Hogues, B. H. Hollins, University of Colorado Anschutz Medical Campus, Denver, CO, USA. 16Cognitive Ethology Laboratory, L. J. Jackson, M. J. Javaid, J. C. Jayaseelan, A. J. Johnson, B. J. Johnson, J. J. Jones, V. J. Joshi, D. Kalra, German Primate Center, Leibniz Institute for Primate Research, Göttingen, Germany. 17Department J. K. Kalu, N. K. Khan, L. Kisamo, L. L. Lago, Y. Lai, F. L. Lara, T.-K. Le, F. L. Legall-Iii, S. L. Lemon, of Biology, University of Konstanz, Konstanz, Germany. 18Centre of Biological Diversity, School of L. Lewis, J. L. Liu, Y.-S. Liu, D. L. Liyanage, P. London, J. L. Lopez, L. L. Lorensuhewa, E. Martinez, Biology, St. Andrews, UK. 19Primate Genetics Laboratory, German Primate Center, Leibniz Institute for R. M. Mata, T. M. Mathew, T. Matskevitch, C. M. Mercado, I. M. Mercado, K. M. Morales, M. M. Morgan, Primate Research, Göttingen, Germany. 20Institute of Evolutionary Biology (UPF-CSIC), PRBB, Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 13 of 14 SCIENCE ADVANCES | RESEARCH ARTICLE Barcelona, Spain. 21School of Public Health, University of Texas Health Science Center, Houston, TX, J. Rogers, C. J. Jolly, D. Zinner, and C. Roos. Cytogenetics and assembly validation: M. Rocchi*, USA. 22Catalan Institution of Research and Advanced Studies (ICREA), Barcelona, Spain. 23CNAG-CRG, R. Stanyon, E. E. Eichler, N. Archidiacono, A. Palazzo, and O. Capozzi. Gene family analysis: Centre for Genomic Regulation, Barcelona Institute of Science and Technology, Barcelona, Spain. M. W. Hahn*, J. Sikela*, G. W. C. Thomas, V. Searles Quick, A. Karimpour-Fard, and L. Walter. 24 Department of Physiology and Biophysics, Weill Cornell Medical College, New York, NY, USA, and Methylation analysis: J. Tung* and T. P. Vilgalys. Positive selection analysis: C. Kosiol*, T. Vinar*, and HRH Prince Alwaleed Bin Talal Bin Abdulaziz Alsaud Institute of Computational Biomedicine, B. J. Raney. Posttranslational modifications: P. Radivojac*, K. A. Pagel, and V. Pejaver. Segmental Weill Cornell Medicine, New York, NY, USA. 25Department of Computer Science and Informatics, duplication analysis: E. E. Eichler*, M. Ventura, A. Raja, C. Catacchio, A. Marra-Campanale, and Indiana University, Bloomington, IN, USA. 26Department of Neuroscience, Washington University J. Huddleston. Copy number variation: T. Marques-Bonet*, L. Kuderna, and M. d. M. Montero. School of Medicine, St. Louis. MO, USA, and Department of Anthropology, Washington University, Transcriptome analysis: C. E. Mason* and L. Pipes. Provided essential biomaterials: K. Rice, Seattle, WA, USA. 27Genomics Institute, University of California, Santa Cruz, CA, USA. 28Institut für J. Pecotte, J. Phillips-Conroy, C. J. Jolly, J. Rogers, J. Else, and L. A. Cox. Provided text and/or Populationsgenetik, Veterinärmedizinische Universität Wien, Vienna, Austria. 29Department of figures: D. Zinner, C. Roos, T. Mailund, K. Leppälä, E. E. Eichler, G. Athanasiadis, J. Cheng, K. Munch, Biology, University of Florence, Florence, Italy. 30Department of Evolutionary Anthropology, C. Kosiol, C. Bergey, A. Burrell, M. K. Konkel, J. A. Walker, M. A. Batzer, and J. Tung. Wrote the paper: Duke University, Durham, NC, USA. 31Faculty of Mathematics, Physics and Informatics, Comenius J. Rogers*, C. J. Jolly, J. Tung, M. Hahn, D. Zinner, C. Roos, T. Marques-Bonet, and K. C. Worley. University, Bratislava, Slovakia. 32Department of Genetics and Biochemistry, Clemson University, *Group leader. Clemson, SC, USA. 33Institut Catala de Paleontologia Miquel Crusafont, Universitat Autonoma de Barcelona, Barcelona, Spain. Submitted 6 July 2018 Consortium Member Contributions: Designed the study and supervised the analysis: J. Rogers*, Accepted 6 December 2018 K. C. Worley, and R. A. Gibbs. Managed or supervised the sequence production: D. M. Muzny*, Published 30 January 2019 C. L. Kovar, H. H. Dinh, and Y. Han. Managed or supervised the preparation of sequencing libraries: 10.1126/sciadv.aau6947 H. Doddapaneni*, S. Lee, and D. M. Muzny. Produced the assembly: K. C. Worley*, Y. Liu, S. Murali, and R. A. Harris. Project and data management: D. M. Muzny*, M. Raveendran, R. A. Harris, Citation: J. Rogers, M. Raveendran, R. A. Harris, T. Mailund, K. Leppälä, G. Athanasiadis, M. H. Schierup, K. C. Worley, S. N. Jhangiani, V. Korchina, C. Kovar. Genome annotation: B. Aken*, F. J. Martin, J. Cheng, K. Munch, J. A. Walker, M. K. Konkel, V. Jordan, C. J. Steely, T. O. Beckstrom, C. Bergey, Downloaded from http://advances.sciencemag.org/ on February 10, 2019 M. Muffato, K. Billis, and X. Liu. Alu repeat analysis: M. A. Batzer*, J. A. Walker, M. K. Konkel, V. E. Jordan, A. Burrell, D. Schrempf, A. Noll, M. Kothe, G. H. Kopp, Y. Liu, S. Murali, K. Billis, F. J. Martin, M. Muffato, C. J. Steely, and T. O. Beckstrom. SNV and indel analysis: R. A. Harris and M. Raveendran. Admixture L. Cox, J. Else, T. Disotell, D. M. Muzny, J. Phillips-Conroy, B. Aken, E. E. Eichler, T. Marques-Bonet, and phylogenetic analysis: T. Mailund, M. H. Schierup, K. Leppälä, J. Cheng, K. Munch, and C. Kosiol, M. A. Batzer, M. W. Hahn, J. Tung, D. Zinner, C. Roos, C. J. Jolly, R. A. Gibbs, K. C. Worley, G. Athanasiadis. Phylogenetic and population analysis: C. Bergey, A. Burrell, A. Noll, D. Schrempf, Baboon Genome Analysis Consortium, The comparative genomics and complex population C. Kosiol, G. H. Kopp, G. Athanasiadis, K. Munch, J. Phillips-Conroy, M. Kothe, T. Disotell, J. Tung, history of Papio baboons. Sci. Adv. 5, eaau6947 (2019). Rogers et al., Sci. Adv. 2019; 5 : eaau6947 30 January 2019 14 of 14 The comparative genomics and complex population history of Papio baboons Jeffrey Rogers, Muthuswamy Raveendran, R. Alan Harris, Thomas Mailund, Kalle Leppälä, Georgios Athanasiadis, Mikkel Heide Schierup, Jade Cheng, Kasper Munch, Jerilyn A. Walker, Miriam K. Konkel, Vallmer Jordan, Cody J. Steely, Thomas O. Beckstrom, Christina Bergey, Andrew Burrell, Dominik Schrempf, Angela Noll, Maximillian Kothe, Gisela H. Kopp, Yue Liu, Shwetha Murali, Konstantinos Billis, Fergal J. Martin, Matthieu Muffato, Laura Cox, James Else, Todd Disotell, Donna M. Muzny, Jane Phillips-Conroy, Bronwen Aken, Evan E. Eichler, Tomas Marques-Bonet, Carolin Kosiol, Mark A. Batzer, Matthew W. Hahn, Jenny Tung, Dietmar Zinner, Christian Roos, Clifford J. Jolly, Richard A. Gibbs, Kim C. Worley and Baboon Genome Analysis Consortium Sci Adv 5 (1), eaau6947. DOI: 10.1126/sciadv.aau6947 Downloaded from http://advances.sciencemag.org/ on February 10, 2019 ARTICLE TOOLS http://advances.sciencemag.org/content/5/1/eaau6947 SUPPLEMENTARY http://advances.sciencemag.org/content/suppl/2019/01/28/5.1.eaau6947.DC1 MATERIALS REFERENCES This article cites 66 articles, 8 of which you can access for free http://advances.sciencemag.org/content/5/1/eaau6947#BIBL PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions Use of this article is subject to the Terms of Service Science Advances (ISSN 2375-2548) is published by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title Science Advances is a registered trademark of AAAS.