lable at ScienceDirect

Journal of Human Evolution 108 (2017) 62e71

Contents lists avai

Journal of Human Evolution

journal homepage: www.elsevier .com/locate/ jhevol

An evolutionary medicine perspective on Neandertal extinction

Alexis P. Sullivan a, Marc de Manuel c, Tomas Marques-Bonet c, d, e, George H. Perry a, b, *

a Department of Biology, Pennsylvania State University, University Park, PA 16802, USAb Department of Anthropology, Pennsylvania State University, University Park, PA 16802, USAc Institut de Biologia Evolutiva (CSIC/UPF), Parque de Investigaci�on Biom�edica de Barcelona (PRBB), Barcelona, Catalonia 08003, Spaind CNAG-CRG, Centre for Genomic Regulation (CRG), Barcelona Institute of Science and Technology (BIST), Baldiri i Reixac 4, 08028 Barcelona, Spaine Catalan Institution of Research and Advanced Studies (ICREA), Passeig de Lluís Companys, 23, 08010, Barcelona, Spain

a r t i c l e i n f o

Article history:Received 12 May 2016Accepted 12 March 2017Available online 16 May 2017

Keywords:Archaic hominin admixturePaleoepidemiologyGenetic driftHuman-pathogen co-evolution

* Corresponding author.E-mail address: ghp3@psu.edu (G.H. Perry).

http://dx.doi.org/10.1016/j.jhevol.2017.03.0040047-2484/© 2017 Elsevier Ltd. All rights reserved.

a b s t r a c t

The Eurasian sympatry of Neandertals and anatomically modern humans e beginning at least 45,000years ago and possibly lasting for more than 5000 years e has sparked immense anthropological interestinto the factors that potentially contributed to Neandertal extinction. Among many different hypotheses,the “differential pathogen resistance” extinction model posits that Neandertals were disproportionatelyaffected by exposure to novel infectious diseases that were transmitted during the period of spatio-temporal sympatry with modern humans. Comparisons of new archaic hominin paleogenome sequenceswith modern human genomes have confirmed a history of genetic admixture e and thus direct contacte between humans and Neandertals. Analyses of these data have also shown that Neandertal nucleargenome genetic diversity was likely considerably lower than that of the Eurasian anatomically modernhumans with whom they came into contact, perhaps leaving Neandertal innate immune systems rela-tively more susceptible to novel pathogens. In this study, we compared levels of genetic diversity ingenes for which genetic variation is hypothesized to benefit pathogen defense among Neandertals andAfrican, European, and Asian modern humans, using available exome sequencing data (three individuals,or six chromosomes, per population). We observed that Neandertals had only 31e39% as many non-synonymous (amino acid changing) polymorphisms across 73 innate immune system genes compared tomodern human populations. We also found that Neandertal genetic diversity was relatively low in anunbiased set of balancing selection candidate genes for primates, those genes with the highest 1% geneticdiversity genome-wide in non-human hominoids (apes). In contrast, Neandertals had similar or higherlevels of genetic diversity than humans in 12 major histocompatibility complex (MHC) genes. Thus, whileNeandertals may have been relatively more susceptible to some novel pathogens and differentialpathogen resistance could be considered as one potential contributing factor in their extinction, theexpectations of this model are not universally met.

© 2017 Elsevier Ltd. All rights reserved.

1. Introduction

Neandertals went extinct ~41e39 ka, following spatiotemporaloverlap with anatomically modern humans in Europe for 2600 to5400 years (Higham et al., 2014), with possibly longer overlap inthe Middle East (Barker et al., 2007; Demeter et al., 2012).Numerous hypotheses to explain Neandertal extinction have beenproposed. Although it has been suggested that climatic fluctuations~40 ka played a key role in the Neandertal extinction process(Tzedakis et al., 2007; Golovanova et al., 2010; Valet and Valladas,

2010), this scenario seems unlikely given the environmental resil-ience demonstrated by Neandertals during previous periods ofintense climate change (Lowe et al., 2012). Most other extinctionhypotheses focus on Neandertalemodern human competition(Horan et al., 2005; Banks et al., 2008; Raichlen et al., 2011;Sandgathe et al., 2011; Gilpin et al., 2016). For example, poten-tially shorter inter-birth intervals for modern humans could haveallowed more rapid population growth compared to Neandertals,facilitating eventual replacement (Trinkaus, 1984; Ponce de Leonet al., 2008). Alternatively, anthropologists have speculated thatthe intelligence and language capabilities of anatomically modernhumans were greater than those of Neandertals (Chase and Dibble,1987; Davidson and Noble, 1989; Marwick, 2003; Maricic et al.,

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e71 63

2013), perhaps facilitating competitive hunting and other subsis-tence strategy advantages through the development of more effi-cient tool technologies (Benazzi et al., 2015; El Zaatari et al., 2016).A recent proposal is that modern humans benefitted from the earlydomestication of dogs, which may have aided large animal hunts toincrease caloric yields for the modern humans and fuel their rapidpopulation growth and ultimately larger population sizes(Shipman, 2015).

Given the important role of disease in population dynamics, ithas also been hypothesized that viral disease transmission frommodern humans could have contributed to the ultimate disap-pearance of the Neandertals (Wolff and Greenwood, 2010). Thisnotion has recently been echoed and expanded upon by Houldcroftand Underdown (2016). Such a “differential pathogen resistance”model would require an anatomically modern human pathogen (orpathogens) of limited virulence for out-of-Africa migrating humanpopulations, but one that would have strongly affected immuno-logically naïve Neandertal populations upon contact and trans-mission. Moreover, the viability of this scenario also requires i)relatively fewer Neandertal pathogens/disease strains at the time ofcontact, ii) some mechanism by which modern humans might nothave been as negatively affected as Neandertals upon novel path-ogen exposure, or iii) both of these factors to be present. Otherwise,post-contact modern human populations would have been equallyadversely affected by exposure to novel Neandertal pathogens. Inthis paper we assess the plausibility of the differential pathogenresistance model by comparing levels of genetic diversity betweenNeandertal and modern human populations in genes for whichgenetic variation is hypothesized to benefit pathogen defense.

Recent advances in genomic sequencing technologies andancient DNA methods have facilitated the generation of a high-quality Altai Neandertal nuclear genome sequence from Siberia(dated to ~50 ka; Green et al., 2010).When analyzed in combinationwith modern human genomic data, this genome has providedconvincing evidence that anatomically modern humans and Ne-andertals interbred, with some introgressed Neandertal haplotypespreserved in non-African modern human populations (Green et al.,2010; Sankararaman et al., 2012; Prüfer et al., 2014; Vernot andAkey, 2015; Sams et al., 2016; Simonti et al., 2016; Nielsen et al.,2017). The requisite intercourse demonstrates at least some levelof direct contact between these populations, and thus opportu-nities for the transfer of infectious diseases. Moreover, analyses ofboth the high-coverage diploid nuclear genome sequence from theAltai Neandertal and mitochondrial DNA sequence data that areavailable for multiple Neandertal individuals suggest that Nean-dertal genetic diversity was substantially lower than that observedwithin modern human populations (Briggs et al., 2009; Green et al.,2010; Dalen et al., 2012; Prüfer et al., 2014). Recently, Castellanoet al. (2014) used a DNA capture method to sequence the exomes(protein-coding regions of the nuclear genome) of two additionalNeandertals, individuals who lived ~49 ka (Wood et al., 2013) and~44 ka (Krings et al., 2000; Green et al., 2010) in Spain and Croatia,respectively. Observed levels of heterozygosity for these two Ne-andertals are also relatively low, suggesting that low nucleargenome genetic diversity was a general Neandertal characteristicrather than restricted to an Altai Neandertal population isolate(Castellano et al., 2014). Specifically, considering only sites withsequence coverage sufficient for single nucleotide polymorphism(SNP) identification for each of the three Neandertals, only 30.3%,44.9%, and 45.3% synonymous SNPs (i.e., those that do not changeamino acids) were observed in Neandertals compared to equal-sized population samples of modern human Africans, Europeans,and Asians, respectively (Castellano et al., 2014).

Within genes directly related to immune function, greaterfunctional genetic diversity increases the potential responsiveness

of the immune system to foreign pathogens (Markert et al., 2004;Wolff and Greenwood, 2010). Balancing selection is thought tomaintain advantageous functional diversity (i.e., nonsynonymous,or amino acid-changing, SNPs) within these genes (Andres et al.,2009; Qutob et al., 2012), and individuals with more genetic di-versity across the genome tend to have higher fitness (Markertet al., 2004). Based on population genetic theory, genetic drift is arelatively stronger force, while natural selection is relatively lesseffective, in smaller populations (Gravel, 2016; Henn et al., 2016).Thus, compared to a larger population, a population with a his-torically small effective population size may have lower geneticdiversity in general across the genome, and different patterns ofdiversity at loci affecting individual health and fitness. Indeed,along with relatively reduced overall genetic diversity, Castellanoet al. (2014) observed a higher proportion of predicted damagingnonsynonymous SNPs than benign nonsynonymous SNPs in Ne-andertals compared to modern humans, consistent with thereduced effectiveness of purifying selection to remove or reducethe frequencies of strongly deleterious variants in Neandertals(Hughes et al., 2003; Zhao et al., 2003; Do et al., 2015; Harris andNielsen, 2016; Juric et al., 2016).

In addition to purifying selection, other types of natural selec-tion, including balancing selection, are also expected to be lesseffective in smaller populations. Thus, the generally low geneticdiversity of Neandertals relative to humans may even be exacer-bated at functional sites in genes related to immune function thatwould otherwise be preserved via balancing selection. Theoreti-cally, such a difference could have facilitated the differentialmorbidity following contact and infectious disease transfer be-tween Neandertals andmodern humans potentially required underthe Pleistocene epidemiological scenarios (the differential path-ogen resistance model) detailed by Wolff and Greenwood (2010)and Houldcroft and Underdown (2016).

In this study we compared the levels and patterns of geneticvariation between Neandertal and modern human populations at i)73 genes associated with innate immune functions, ii) 164 virus-interacting protein genes, iii) 12 major histocompatibility com-plex (MHC) genes, and iv) the 1% of genes across the genome withthe consistently highest levels of genetic diversity among four apespecies. Our analysis represents an evaluation of the plausibility ofthe differential pathogen resistance model as a factor potentiallycontributing to Neandertal extinction. Specifically, relatively lowergenetic variation in Neandertal populations among genes in thesefour categories would be consistent with the idea of greater sus-ceptibility to novel pathogens in Neandertals compared to themodern human populations with which they interacted. Incontrast, similar levels of Neandertal and modern human geneticdiversity would raise major questions about the plausibility of thisepidemiological extinction hypothesis.

2. Materials and methods

We downloaded the Neandertal exome DNA capture data pub-lished by Castellano et al. (2014) (http://cdna.eva.mpg.de/neandertal/exomes/VCF). Specifically, we considered the SNP ge-notype data for autosomal chromosomes from the “combined” VCFfiles from this dataset, in which SNP genotypes for each of the 13individuals in the dataset (three modern humans of Africandescent, three modern humans of European descent, three modernhumans of Asian descent, three Neandertal individuals, and oneindividual from the archaic hominin Denisovan population) wereprovided for only the individuals with a minimum of six inde-pendent sequencing reads at that position. The nine modern hu-man individuals and three Neandertal individuals included in theCastellano et al. (2014) dataset are listed in Table 1. Ancient DNA

Table 1Nine modern human and three Neandertal individuals whose genome data were analyzed in this study.

Population Sampling location (population) Individual ID Years before present Original data source

Modern human e Africa Nigeria (Yoruba) HGDP00927 Modern (Meyer et al., 2012)Modern human e Africa Senegal (Mandenka) HGDP01284 Modern (Meyer et al., 2012)Modern human e Africa Sudan (Dinka) DNK2 Modern (Meyer et al., 2012)Modern human e Europe France (French) HGDP00521 Modern (Meyer et al., 2012)Modern human e Europe Italy (Sardinian) HGDP00665 Modern (Meyer et al., 2012)Modern human e Europe USA (Italian American) NA12891 Modern (Lao et al., 2008)Modern human e Asia China (Han) HGDP00778 Modern (Meyer et al., 2012)Modern human e Asia China (Dai) HGDP01307 Modern (Meyer et al., 2012)Modern human e Asia New Guinea (Papuan) HGDP00542 Modern (Meyer et al., 2012)Neandertal Altai, Siberia Toe bone ~50,000 (Prüfer et al., 2014)Neandertal El Sidr�on, Spain SD1253 ~49,000 (Castellano et al., 2014)Neandertal Vindija, Croatia Vi33.15 ~44,000 (Castellano et al., 2014)

Figure 1. Genome-wide SNP derived allele frequency distributions for Neandertalsand three modern human populations. The number of SNPs observed for each popu-lation at exome sites with sufficient sequence coverage for SNP identification in allindividuals in the study, binned by derived allele frequency. A) Synonymous SNPs. B)Nonsynonymous SNPs, with “predicted benign” and “predicted damaging” SNPsindicated separately (lighter and heavier shades respectively), based on PolyPhen-2(Adzhubei et al., 2010) estimates. (For interpretation of the references to color inthis figure legend, the reader is referred to the web version of this article.)

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e7164

data of comparable sequence coverage for three Late Pleistoceneanatomically modern humans who lived contemporaneously withNeandertals (or, more nearly contemporaneously with Neandertalsthan the modern human genomes represented in the Castellanoet al. [2014] dataset) are not yet available.

Genotypes were estimated by Castellano et al. (2014) for onlythose positions covered by a minimum of six independentsequencing reads; the authors also tested higher minimumcoverage cutoffs (up to 20x) that did not significantly affect theobserved patterns of Neandertal versus modern human geneticdiversity, so the lower coverage cutoff was used to include a largernumber of sites in the analysis (Castellano et al., 2014). For ouranalysis, we considered only the autosomal sites with sufficientcoverage for SNP genotyping among all individuals, so that thenumbers of identified SNPs per population and their allele fre-quency distributions could be compared directly among theNeandertal and modern human African, European, and Asiansamples (n ¼ 6 chromosomes per population) as measures of ge-netic diversity.

There were a total of 69,230 autosomal SNPs in the dataset. Weremoved n ¼ 32,416 SNPs for which genotypes were not estimatedfor all of the Neandertal and modern human individuals and n ¼ 17SNPs with more than two identified alleles (e.g., A/T/C variablypresent at one position) (Castellano et al., 2014). The remaining36,797 autosomal SNPs were submitted to PolyPhen-2's onlineHumDiv server via batch query for identification of non-synonymous and synonymous SNPs and estimates of predictedfunctional consequences for the nonsynonymous SNPs (Adzhubeiet al., 2010; Castellano et al., 2014).

PolyPhen-2 classified each SNP as “missense” (non-synonymous), “coding-synon” (synonymous), “nonsense” (stopcodon), “utr-5” (50 untranslated region), “utr-3” (30 untranslatedregion), “intron” (non-coding variants). Our subsequent analysisfocused only on the nonsynonymous (n ¼ 16,139) and synonymous(n ¼ 18,095) SNPs, and excluded stop codons and untranslatedregions. For each nonsynonymous SNP, PolyPhen-2 also provided“benign” (n ¼ 10,558), “possibly damaging” (n ¼ 2222), and“probably damaging” (n¼ 3317) predictions (Adzhubei et al., 2010).PolyPhen-2 failed to make functional predictions for a smallnumber of nonsynonymous SNPs (n ¼ 42), which were removedfrom our subsequent analyses. Single nucleotide polymorphismspredicted to be in the possibly and probably damaging categorieswere combined into one “predicted damaging” category for ouranalyses. PolyPhen-2 assembled gene names from the UCSCknownGene transcripts/database (hg19/GRCh37), and these genenames were used for later identification. In our final, curateddataset, 11,299 genes were represented by at least one SNP. Theannotated database of nonsynonymous and synonymous SNP ge-notypes and population frequencies that we analyzed for this study

are available in Supplementary Online Material [SOM] Database 1(https://scholarsphere.psu.edu/files/v405s943f). All analyses ofthe SNP genotype data were performed using the R statisticalenvironment (R Core Team, 2015).

In addition, to compare patterns of synonymous and non-synonymous SNP derived allele frequencies among populations(e.g., see Fig. 1) we determined the derived and ancestral states for

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e71 65

the alleles of each SNP by comparison to the orthologous chim-panzee, gorilla, and orangutan nucleotides based on alignments toreference genomes for those species as provided by Castellano et al.(2014). Ancestral alleles were classified as those that matched theorthologous nucleotides for all three ape species. We removed fromthe dataset SNPs with i) variability, ii) missing data, or iii) nucleo-tides different than either of the two Neandertal/human alleles atthe orthologous positions among the three ape species (n ¼ 538nonsynonymous and n ¼ 738 synonymous SNPs) in our analyses ofderived allele frequencies. Note that our comparisons of the pat-terns of genetic diversity among gene categories (e.g., see Figs. 2and 3) did not require derived allele frequency information; thus,the SNPs with variability, missing data, or nucleotides differentthan the two Neandertal/human alleles were still included in thatanalysis to avoid bias against highly variable loci (e.g., those withcross-species polymorphisms maintained by long-term balancingselection).

A potential confounding factor in our analysis is the effect ofNeandertal to human introgression on patterns of genetic diversity,especially considering that surviving introgressed loci may beenriched for regions of the genome containing genes involved inimmune processes (Dannemann et al., 2016; Sams et al., 2016;Racimo et al., 2017). Genetic diversity in African modern humanpopulations would be unaffected by this history of introgression.For non-African modern human populations any introgression ef-fect would be in one direction only (an increase in diversity). Wecannot completely control or account for this effect, which gives usreason to be cautious when interpreting results. Meanwhile, theAltai Neandertal individual's genome is reported to include someDNA segments that were introgressed from modern humans toNeandertals (Kuhlwilm et al., 2016), which could inflate the overallNeandertal genetic diversity estimates. It was possible to assess thepotential effect of this pattern of introgression by repeating ana-lyses in which Neandertal genetic diversity was relatively high,following the removal of this individual from the dataset.

We evaluated patterns of Neandertal-modern human geneticdiversity within three subsets of genes for which genetic diversityitself is thought to play an important role in pathogen defense-related immune functions. The first set comprised 73 innate im-mune receptor, signaling adaptor molecule, and complementpathway genes (“Innate Immune SystemGenes”; SOM Table S1), forexample the toll-like receptor (Akira et al., 2001; Ferrer-Admetllaet al., 2008; Netea and Joosten, 2016) and mannose-associatedserine protease genes (Fujita, 2002). Many of the proteins enco-ded by these genes are involved in the innate immune system's firstline of defense against diverse external microorganisms. The sec-ond set comprised 164 genes encoding for virus-interacting pro-teins that also have known antiviral activity or broader roles in theimmune system (Enard et al., 2016; SOM Table S2). The third setcomprised 12 MHC genes (SOM Table S3), critical immune systemloci with among the strongest evidence for long-term balancingselection and long-term maintenance of allelic diversity in verte-brate genomes (Loisel et al., 2006; Leffler et al., 2013; Lenz et al.,2013; Azevedo et al., 2015). All available genes for each of thesegene sets from the Castellano et al. (2014) database were analyzed,except for theMHC pseudogene HLA-DPB2 (Gustafsson et al., 1987),which we excluded. The patterns of Neandertal andmodern humangenetic diversity for these three gene sets were compared to thosefor the remaining 11,086 genes in our genome-wide dataset.

We also evaluated patterns of Neandertal-modern human ge-netic diversity within a set of genes with consistently high levels ofgenetic diversity among apes. To avoid introducing bias into thisNeandertal versus human comparison, we did not use existingbalancing selection candidate gene lists, because the majority ofpublished studies of balancing selection in mammals have been

based at least in part on human population genomic data (Andreset al., 2009; Leffler et al., 2013; DeGiorgio et al., 2014; Key et al.,2014; Rasmussen et al., 2014; Azevedo et al., 2015; Gao et al.,2015). Instead, we analyzed genome-wide sequence data from 55total individuals from population samples of four ape species toidentify an unbiased (with respect to our analysis) set of candidatebalancing selection genes.

Specifically, Prado-Martinez et al. (2013) originally producedsequence data for 87 great ape individuals, with DNA from eachindividual sequenced on an Illumina platform to ~25x sequencecoverage. Reads were mapped to the human reference genome(hg18). Genotypeswere estimated only for sites that met criteria forsequence coverage, base quality, and mapping quality (“callablesites”), as described by Prado-Martinez et al. (2013). For our geneticdiversity analysis, we focused on a subset of these data (n ¼ 55individuals), considering the one population from each ape speciesor species group with the largest sample size: chimpanzee (Pantroglodytes ellioti; n ¼ 10), bonobo (Pan paniscus; n ¼ 13), gorilla(Gorilla gorilla gorilla; n ¼ 27), and orangutan (Pongo abelii; n ¼ 5).We obtained hg18 gene coding region coordinates from theknownCanonical gene database using the table browser at the UCSCGenome Bioinformatics Site (genome.ucsc.edu/index.html). Foreach gene also in the Neandertal-human genetic diversity databasewe estimated nucleotide diversity as the average proportion ofpairwise differences (p) for the callable sites of the coding regions(Nei and Li, 1979). There were 7259 genome-wide genes with�500“callable” sites across all four species for which there was also atleast one variable site in our Neandertal-modern human geneticdiversity database. Within each species, we computed p valuepercentiles for each gene, and then summed the per-genepercentile values across species. Using this approach, we identi-fied the 73 (top 1%) genes with the consistently highest geneticdiversity among these ape species (“top 1% ape diversity genes”;SOM Table S4). The great ape gene diversity data analyzed in thisstudy are available in SOM Database 2 (https://scholarsphere.psu.edu/files/5d86p0267). Gene Ontology enrichment analyses wereperformed using the WEB-based Gene SeT AnaLysis Toolkit (Web-Gestalt). Functional category enrichments were statistically evalu-ated with hypergeometric tests, and p-values were adjusted formultiple tests using the method of Benjamini and Hochberg (1995).

For each of the four different gene categories considered in thisstudy, our analysis included every locus represented with one ormore nonsynonymous or synonymous SNPs in the Castellano et al.(2014) dataset, excepting the MHC pseudogene HLA-DPB2(Gustafsson et al., 1987), which we excluded from the MHC geneset (SOM Tables S1, S2, S3, S4).

3. Results

We analyzed a combined ancient and modern genome exomeSNP dataset that was originally produced by Castellano et al. (2014)in order to compare levels and patterns of genetic diversity be-tween Neandertals andmodern humans as part of an assessment ofthe plausibility of the differential pathogen resistance model toexplain the extinction of Neandertals.

3.1. Patterns of genetic diversity and the effectiveness of selection

On the overall dataset, we observed patterns of genetic diversitythat were similar to the primary findings of Castellano et al. (2014).Specifically, Neandertals had relatively fewer total SNPs than any ofthe modern human populations; this was true for both non-synonymous and synonymous SNPs (Fig. 1; SOM Table S5). Inaddition, Neandertals also had a higher proportion of non-synonymous to total (nonsynonymous þ synonymous) SNPs

Figure 2. Patterns of Neandertal and modern human nonsynonymous (nonsyn.) SNP diversity genome-wide, in innate immune genes, in virus-interacting protein genes, and inMHC genes. The number of nonsynonymous SNPs per population for sites with sufficient sequence coverage for SNP identification in all individuals in the study, with PolyPhen-2“predicted benign” (lighter shading) and “predicted damaging” (heavier shading) SNPs indicated separately, and the proportion of the nonsynonymous SNPs with minor allelefrequencies ¼ 1, 2, and 3 (of n ¼ 6 total chromosomes per population). A and B) Nonsynonymous SNPs in the genome-wide set of n ¼ 11,086 genes that excludes innate immunereceptor, signaling adaptor molecule, complement pathway, virus-interacting protein, and MHC genes. C and D) Nonsynonymous SNPs in the set of n ¼ 73 innate immune receptor,signaling adaptor molecule, and complement pathway genes. E and F) Nonsynonymous SNPs in the set of n ¼ 164 virus-interacting protein genes with known antiviral or immuneactivities. G and H) Nonsynonymous SNPs in the set of n ¼ 12 MHC genes. MAF ¼ minor allele frequency; colors in Figure 2B, D, F, H follow key on bar charts. (For interpretation ofthe references to color in this figure legend, the reader is referred to the web version of this article.)

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e7166

Figure 3. Patterns of Neandertal and modern human nonsynonymous (nonsyn.) SNP diversity genome wide and for genes with high genetic diversity in non-human hominoids(apes). The number of nonsynonymous SNPs per population for sites with sufficient sequence coverage for SNP identification in all individuals in the study, with PolyPhen-2“predicted benign” (lighter shading) and “predicted damaging” (heavier shading) SNPs indicated separately, and the proportion of the nonsynonymous SNPs with minor allelefrequencies ¼ 1, 2, and 3 (of n ¼ 6 total chromosomes per population). A and B) Nonsynonymous SNPs in the genome-wide set of n ¼ 7186 genes with at least 500 “callable” siteswith sufficient coverage and mapping quality for SNP identification in a great ape population genomics panel, excluding the top 1% highest diversity genes among four ape species. Cand D) Nonsynonymous SNPs in the set of n ¼ 73 genes with the consistently highest levels of genetic diversity among the four ape species. MAF ¼ minor allele frequency; colors inFigure 3B, D follow key on bar charts. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e71 67

compared to the human populations (Neandertals ¼ 0.500; humanpopulations ¼ 0.440e0.452; Fisher's Exact Test for the Neandertalversus European comparison; p ¼ 2.29 � 10�10; see SOM Table S5for all comparisons) and a considerably higher proportion ofpredicted damaging to total nonsynonymous SNPs(Neandertal¼ 0.435; human¼ 0.291e0.297; p < 2.2� 10�16 for theNeandertal-European comparison; SOM Table S6). Together, theseresults are consistent with the notion that purifying natural se-lection against potentially damaging nonsynonymous SNPs wasrelatively less effective in smaller Neandertal populations.

3.2. Innate immune system, virus-interacting protein, and MHCgene diversity comparisons

We next performed analyses focused on patterns of genetic di-versity within three subsets of genes for which genetic diversityitself is thought to play an important role in pathogen defense-related immune functions: 73 innate immune system genes (SOMTable S1), 164 genes encoding for virus-interacting proteins thatalso have known antiviral activity or broader roles in the immunesystem (Enard et al., 2016; SOM Table S2), and 12 MHC genes (SOMTable S3). Patterns of Neandertal and modern human genetic di-versity for these three gene sets were compared to those for theremaining 11,086 genes in our genome-wide dataset.

Neandertal and human nonsynonymous genetic diversity wassimilar between the genome-wide and innate immune systemgene sets, with fewer nonsynonymous SNPs in Neandertals

compared to any human population in both cases (Fig. 2AeD). Infact, the relative number of Neandertal versus human non-synonymous SNPs is even slightly higher for the genome-wideset than for the innate immune system genes; e.g., with 54.1%as many Neandertal as European human nonsynonymous SNPs inthe genome-wide set compared to only 39.0% for the innateimmune genes. The direction of this result is consistent withexpectations under a model of relatively reduced balancing se-lection effectiveness in Neandertals (i.e., if innate immune genefunctional genetic diversity confers a fitness advantage). How-ever, the observed Neandertal-human difference between thetwo gene sets is not significant based on Fisher's Exact Tests(p ¼ 0.33 for Neandertals-Europeans; see SOM Table S7 for allcomparisons). Based on the results of a permutation analysis with10,000 sets of 73 random genes from the genome-wide set, theproportion of the number of Neandertal-human nonsynonymousSNPs is also not significantly lower than expected by chance(p ¼ 0.1944 for Neandertals-Europeans; see SOM Figure S1 for allcomparisons). Regardless, our results do not provide any supportfor the possibility that strong balancing selection has maintainedsimilar levels of nonsynonymous diversity in Neandertal andhuman innate immune system genes despite the lower effectivepopulation size and genome-wide diversity of Neandertals. Therelative number of Neandertal versus human nonsynonymousSNPs observed among the virus-interacting protein genes wasalso low (Fig. 2EeF), in this case at similar proportions to thegenome-wide gene set (SOM Table S8; SOM Figure S2).

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e7168

In contrast, there was a greater number of total nonsynonymousSNPs observed across the 12 MHC genes for Neandertals (17) thanfor any of the three human populations (9, 13, and 9 for the African,European, and Asian population samples, respectively; Fig. 2G).Compared to both the genome-wide set of genes and the innateimmune system genes, the number of Neandertal relative to humanMHC gene nonsynonymous SNPs is significantly greater than ex-pected by chance when assessed with Fisher's Exact Tests (e.g.,Neandertals-Europeans forMHC versus genome-wide set; p¼ 0.02;see SOM Table S9 for all comparisons) and, for some but not allpopulation comparisons, with permutation analyses (p¼ 0.1570 forNeandertals-Europeans; p ¼ 0.0887 for Neandertals-Asians;p ¼ 0.0234 for Neandertals-Africans; SOM Figure S3).

Additionally, we observed 9/17 (52.9%) of the Neandertal MHCnonsynonymousSNPs at intermediate frequency (i.e.,with theminorallele observed on three out of the six chromosomes in the popula-tion; Fig. 2H, a significantly higher proportion than observed forNeandertal nonsynonymous SNPs in the genome-wide, non-im-mune gene set (260/2527; 10.3%; Fisher's Exact Test; p¼ 1.56� 10�5;SOM Table S10). The proportion of Neandertal MHC intermediatefrequency nonsynonymous SNPs was also relatively higher for Ne-andertals than for any of the human populations, although with thesmall sample sizes not all comparisons were statistically significant(Fisher's Exact Tests; p ¼ 0.43 for Neandertals-Africans; p ¼ 0.02 forNeandertals-Europeans; p ¼ 0.09 for Neandertals-Asians; SOMTable S11). The Neandertal-human MHC genetic diversity patternsremained similar after we excluded one random individual fromeach human population and the Altai Neandertal individual, whosegenomic MHC region may harbor some DNA segments introgressedfrom humans (Kuhlwilm et al., 2016; SOM Figure S4).

In combination, the relatively large number of Neandertal MHCnonsynonymous SNPs and the high proportion of those variantsobserved at intermediate frequencies suggest that at least for thesecritical immune loci, the heterozygous fitness advantage for Ne-andertals was sufficiently strong to offset the lower effective size ofthis population, leading to similar or even higher putatively func-tional diversity compared to modern human populations.

3.3. Balancing selection candidate gene comparisons

Finally, we sought to compare patterns of Neandertal-humangenetic diversity across genes for which evidence of balancing se-lection has been identified without respect to gene function. Spe-cifically, we identified the 73 (top 1%) genes with the consistentlyhighest genetic diversity among these ape species (“top 1% ape di-versity genes”; SOM Table S4). Relative to all 7259 genes andfollowing correction for multiple tests, the top 1% ape diversity setwas significantly enriched for genes with immune system-relatedGene Ontology functional categories including “MHC protein com-plex” (observed ¼ 3 genes; expected ¼ 0.09 genes; adjustedp ¼ 0.0015), “positive regulation of leukocyte activation”(observed ¼ 6 genes; expected ¼ 0.62 genes; adjusted p ¼ 0.0061),and “immune system process” (observed ¼ 14 genes;expected ¼ 4.92 genes; adjusted p ¼ 0.0098). Full results from theGene Ontology enrichment analysis are provided in SOM Table S12(see also SOM Figure S5). In addition to the MHC loci, the top 1%ape diversity set contains OAS1, another gene in which high allelicdiversity appears likely to have been maintained across multiplespecies by long-term balancing selection (Ferguson et al., 2012).Together, these features suggest that our top 1% ape diversity genesare likely enriched for those affected by balancing selection inhominoid primates. Thus, it is appropriate to compare Neandertaland modern human genetic diversity at these loci as an additionalcomponent of our assessment of the underlying mechanics of thedifferential pathogen resistance model for Neandertal extinction.

Across the top 1% ape diversity genes, we observed only 27%e58% as many nonsynonymous SNPs for Neandertals compared tothe modern human populations (Fig. 3C). The magnitude of thisdifference was similar to that for the remaining genome-widegenes (Fig. 3A; Fisher's Exact Test; p ¼ 0.69 for Neandertals-Europeans; see SOM Table S13 for all comparisons). Thus, thepattern of Neandertal versus modern human nonsynonymous ge-netic diversity for the top 1% ape diversity gene set was moresimilar to that observed for the innate immune system genes thanthe MHC genes.

4. Discussion

The Neandertal and anatomically modern human lineagesdiverged ~550 ka (Krings et al., 1997; Green et al., 2008; Prüfer et al.,2014). Over several hundred thousand years of subsequent homininpopulation isolation, pathogens that originally infected theNeandertal-modern human common ancestor would also haveevolved separatelyand in isolation,withpotential co-evolutionof theimmune systems of each hominin population. Proto-Neandertal andproto-anatomically modern human populations may likewise haveseparately adapted to any pathogens that they newly encountered(and that did not infect the other population) during this period.

These separate pathogen and immune system histories couldhave created the conditions for epidemic disease outbreaks andmortality in immunologically-naïve populations, in the event ofsubsequent direct contact and disease transmission. While thespatiotemporal overlap between Neandertals and modern humansmay have persisted for as few as five thousand years, or even less, insome parts of Europe (Adler et al., 2008; Higham et al., 2014),Neandertal-modern human overlap in the Middle East was likelyconsiderably longer than that in Europe, with an earlier presence ofmodern human fossils (Grün et al., 2005) and solid inference thatgenetic admixture e and thus, direct contact with the opportunityfor infectious disease transfer e had occurred by ~55 ka (Fu et al.,2014; Seguin-Orlando et al., 2014).

At multiple points in modern human history, migration-basedexposures to novel infectious diseases/strains likely contributedto substantial population loss (Perry and Fetherston, 1997;Thornton, 1997; Ramenofsky et al., 2003; Morelli et al., 2010),over much shorter timeframes than the spatiotemporal overlap ofEuropean Neandertals andmodern humans. It should also be notedthat in those examples from modern human history, the pop-ulations had been evolving separately for substantially less timethan humans and Neandertals. Thus, at least theoretically, theprocess of infectious disease exchange between Neandertals andmodern humans during their period of Middle Eastern and Euro-pean spatiotemporal overlap could have resulted in substantialmortality. In this context, any Neandertal-modern human pathogensusceptibility difference theoretically could have affected thecompeting demographic trajectories of these contemporaneouspopulations, which is exactly the thesis of the differential pathogenresistance model of Neandertal extinction. In other words, thishypothesis retrodicts that Neandertals may have been dispropor-tionately affected by transferred infectious diseases after contactwith anatomically modern humans whomigrated out of Africa, to adegree that could have played a role in the ultimate disappearanceof Neandertals.

Our goal in this studywas to assess the underlying plausibility ofthe differential pathogen resistance model by testing whetherNeandertal genetic diversity at critical immune system loci wassignificantly lower than that for the modern human populationswith whom they came into contact. Neandertals apparently didhave lower overall genetic diversity than anatomically modernhumans (Burbano et al., 2010; Castellano et al., 2014; Parks et al.,

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e71 69

2015), and we asked whether genetic diversity at Neandertal im-mune system loci was correspondingly low, which could haveserved as one mechanism for relatively reduced pathogen resis-tance in Neandertals. We note that the differential pathogenresistance model would not require Neandertals to be depauperateof distinct and evolutionarily advantageous genetic variants atimmune system loci. Evidence of the adaptive introgression ofNeandertal toll-like receptor (TLR), oligoadenylate synthetase(OAS), and MHC gene variants into modern human populationsstrongly suggests that Neandertals did indeed possess such alleles(Abi-Rached et al., 2011; Dannemann et al., 2016; Deschamps et al.,2016; Sams et al., 2016). Rather, the mechanics of the differentialpathogen resistance model could still be relevant if there weresimply fewer such variants and substantially lower functional ge-netic diversity overall at immune system loci in Neandertalscompared to sympatric modern human populations.

Are our results consistent with the differential pathogen resis-tance model for Neandertal extinction? Not in full. While Nean-dertal nonsynonymous genetic diversity at innate immune system,virus-interacting protein, and ape high-diversity gene loci wasindeed low compared to that observed for humans, NeandertalMHC diversity was similar or even higher than that for humans inmultiple respects. Based on the relatively high Neandertal MHCdiversity result, future models that incorporate epidemiologicalmechanisms as contributing factors to Neandertal extinctionshould proceed with caution, as there are specific genes for whichbalancing selection in Neandertals appears to have overcome thelower effective size and lower levels of genome-wide genetic di-versity in this population. The absence of a major difference ingenetic diversity acrossMHC loci between Neandertals andmodernhumans does not mean that there might not have been an aggre-gate differential pathogen resistance effect from lower Neandertalfunctional diversity across innate immune system and other criticalpathogen defense genes apart from MHC. In the future, the po-tential applicability of the differential pathogen resistance hy-pothesis as a component of the Neandertal extinction process couldbe explored further with the combination of expanded Neandertaland modern human paleogenomic population data and broadexperimental/functional comparisons of Neandertal versus modernhuman innate immune gene diversity.

Acknowledgments

We thank Sergi Castellano and Luis Barreiro for their discussionson this project, and David Enard and Martin Kuhlwilm for helpfulanalytical suggestions. Funding: The computational resourceinstrumentation used in this study was funded by the NationalScience Foundation (OCIe0821527) and this material is based onwork supported by the National Science Foundation GraduateResearch Fellowship Program (DGE1255832, to A.P.S.).

Supplementary Online Material

Supplementary online material related to this article can befound at http://dx.doi.org/10.1016/j.jhevol.2017.03.004.

References

Abi-Rached, L., Jobin, M.J., Kulkarni, S., McWhinnie, A., Dalva, K., Gragert, L.,Babrzadeh, F., Gharizadeh, B., Luo, M., Plummer, F.A., Kimani, J., Carrington, M.,Middleton, D., Rajalingam, R., Beksac, M., Marsh, S.G.E., Maiers, M.,Guethlein, L.A., Tavoularis, S., Little, A.-M., Green, R.E., Norman, P.J., Parham, P.,2011. The shaping of modern human immune systems by multiregionaladmixture with archaic humans. Science 334, 89e94.

Adler, D.S., Bar-Yosef, O., Belfer-Cohen, A., Tushabramishvili, N., Boaretto, E.,Mercier, N., Valladas, H., Rink, W.J., 2008. Dating the demise: Neandertal

extinction and the establishment of modern humans in the southern Caucasus.J. Hum. Evol. 55, 817e833.

Adzhubei, I.A., Schmidt, S., Peshkin, L., Ramensky, V.E., Gerasimova, A., Bork, P.,Kondrashov, A.S., Sunyaev, S.R., 2010. A method and server for predictingdamaging missense mutations. Nature Methods 7, 248e249.

Akira, S., Takeda, K., Kaisho, T., 2001. Toll-like receptors: critical proteins linkinginnate and acquired immunity. Nature Immunol. 2, 675e680.

Andres, A.M., Hubisz, M.J., Indap, A., Torgerson, D.G., Degenhardt, J.D., Boyko, A.R.,Gutenkunst, R.N., White, T.J., Green, E.D., Bustamante, C.D., Clark, A.G.,Nielsen, R., 2009. Targets of balancing selection in the human genome. Mol.Biol. Evol. 26, 2755e2764.

Azevedo, L., Serrano, C., Amorim, A., Cooper, D.N., 2015. Trans-species poly-morphism in humans and the great apes is generally maintained by balancingselection that modulates the host immune response. Hum. Genomics 9, 21.

Banks,W.E., D'Errico, F., Peterson, A.T., Kageyama,M., Sima, A., S�anchez-Go~ni, M.-F., 2008. Neanderthal extinction by competitive exclusion. PLoS ONE 3,e3972.

Barker, G., Barton, H., Bird, M., Daly, P., Datan, I., Dykes, A., Farr, L., Gilbertson, D.,Harrisson, B., Hunt, C., Higham, T., Kealhofer, L., Krigbaum, J., Lewis, H.,McLaren, S., Paz, V., Pike, A., Piper, P., Pyatt, B., Rabett, R., Reynolds, T., Rose, J.,Rushworth, G., Stephens, M., Stringer, C., Thompson, J., Turney, C., 2007. The“human revolution” in lowland tropical Southeast Asia: the antiquity andbehavior of anatomically modern humans at Niah Cave (Sarawak, Borneo).J. Hum. Evol. 52, 243e261.

Benazzi, S., Slon, V., Talamo, S., Negrino, F., Peresani, M., Bailey, S.E., Sawyer, S.,Panetta, D., Vicino, G., Starnini, E., Mannino, M.A., Salvadori, P.A., Meyer, M.,P€a€abo, S., Hublin, J.-J., 2015. The makers of the Protoaurignacian and implica-tions for Neandertal extinction. Science 348, 793e796.

Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate: A practicaland powerful approach to multiple testing. J. R. Stat. Soc. Series B Stat. Meth-odol. 57, 289e300.

Briggs, A.W., Good, J.M., Green, R.E., Krause, J., Maricic, T., Stenzel, U., Lalueza-Fox, C.,Rudan, P., Brajkovic, D., Kucan, Z., Gusic, I., Schmitz, R., Doronichev, V.B.,Golovanova, L.V., de la Rasilla, M., Fortea, J., Rosas, A., P€a€abo, S., 2009. Targetedretrieval and analysis of five Neandertal mtDNA genomes. Science 325,318e321.

Burbano, H.A., Hodges, E., Green, R.E., Briggs, A.W., Krause, J., Meyer, M., Good, J.M.,Maricic, T., Johnson, P.L.F., Xuan, Z., Rooks, M., Bhattacharjee, A., Brizuela, L.,Albert, F.W., de la Rasilla, M., Fortea, J., Rosas, A., Lachmann, M., Hannon, G.J.,P€a€abo, S., 2010. Targeted investigation of the Neandertal genome by array-based sequence capture. Science 328, 723e725.

Castellano, S., Parra, G., Sanchez-Quinto, F.A., Racimo, F., Kuhlwilm, M., Kircher, M.,Sawyer, S., Fu, Q., Heinze, A., Nickel, B., Dabney, J., Siebauer, M., White, L.,Burbano, H.A., Renaud, G., Stenzel, U., Lalueza-Fox, C., de la Rasilla, M., Rosas, A.,Rudan, P., Brajkovi, D., Kucan, E., Gu ic, I., Shunkov, M.V., Derevianko, A.P.,Viola, B., Meyer, M., Kelso, J., Andres, A.M., P€a€abo, S., 2014. Patterns of codingvariation in the complete exomes of three Neandertals. Proc. Natl. Acad. Sci.U.S.A. 111, 6666e6671.

Chase, P.G., Dibble, H.L., 1987. Middle Paleolithic symbolism: A review of currentevidence and interpretations. J. Anthropol. Archaeol. 6, 263e296.

Dalen, L., Orlando, L., Shapiro, B., Brandstrom-Durling, M., Quam, R., Gilbert, M.T.P.,Diez Fernandez-Lomana, J.C., Willerslev, E., Arsuaga, J.L., Gotherstrom, A., 2012.Partial genetic turnover in Neandertals: Continuity in the East and populationreplacement in the West. Mol. Biol. Evol. 29, 1893e1897.

Dannemann, M., Andr�es, A.M., Kelso, J., 2016. Introgression of Neandertal- andDenisovan-like haplotypes contributes to adaptive variation in human toll-likereceptors. Am. J. Hum. Genet. 98, 22e33.

Davidson, I., Noble, W., 1989. The archaeology of perception: Traces of depiction andlanguage. Curr. Anthropol. 30, 125e155.

DeGiorgio, M., Lohmueller, K.E., Nielsen, R., 2014. A model-based approach foridentifying signatures of ancient balancing selection in genetic data. PLOSGenet. 10, e1004561.

Demeter, F., Shackelford, L.L., Bacon, A.-M., Duringer, P., Westaway, K.,Sayavongkhamdy, T., Braga, J., Sichanthongtip, P., Khamdalavong, P., Ponche, J.-L., Wang, H., Lundstrom, C., Patole-Edoumba, E., Karpoff, A.-M., 2012.Anatomically modern human in Southeast Asia (Laos) by 46 ka. Proc. Natl. Acad.Sci. 109, 14375e14380.

Deschamps, M., Laval, G., Fagny, M., Itan, Y., Abel, L., Casanova, J.-L., Patin, E.,Quintana-Murci, L., 2016. Genomic signatures of selective pressures andintrogression from archaic hominins at human innate immunity genes. Am. J.Hum. Genet. 98, 5e21.

Do, R., Balick, D., Li, H., Adzhubei, I., Sunyaev, S., Reich, D., 2015. No evidence thatselection has been less effective at removing deleterious mutations in Euro-peans than in Africans. Nature Genet. 47, 126e131.

El Zaatari, S., Grine, F.E., Ungar, P.S., Hublin, J.-J., 2016. Neandertal versus modernhuman dietary responses to climatic fluctuations. PLOS ONE 11, e0153277.

Enard, D., Cai, L., Gwennap, C., Petrov, D.A., 2016. Viruses are a dominant driver ofprotein adaptation in mammals. eLife 5, e12469.

Ferguson, W., Dvora, S., Fikes, R.W., Stone, A.C., Boissinot, S., 2012. Long-termbalancing selection at the antiviral gene OAS1 in Central African chimpanzees.Mol. Biol. Evol. 29, 1093e1103.

Ferrer-Admetlla, A., Bosch, E., Sikora, M., Marques-Bonet, T., Ramirez-Soriano, A.,Muntasell, A., Navarro, A., Lazarus, R., Calafell, F., Bertranpetit, J., Casals, F., 2008.Balancing selection is the main force shaping the evolution of innate immunitygenes. J. Immunol. 181, 1315e1322.

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e7170

Fu, Q., Li, H., Moorjani, P., Jay, F., Slepchenko, S.M., Bondarev, A.A., Johnson, P.L.F.,Aximu-Petri, A., Prüfer, K., de Filippo, C., Meyer, M., Zwyns, N., Salazar-García, D.C., Kuzmin, Y.V., Keates, S.G., Kosintsev, P.A., Razhev, D.I., Richards, M.P.,Peristov, N.V., Lachmann, M., Douka, K., Higham, T.F.G., Slatkin, M., Hublin, J.-J.,Reich, D., Kelso, J., Viola, T.B., P€a€abo, S., 2014. Genome sequence of a 45,000-year-old modern human from western Siberia. Nature 514, 445e449.

Fujita, T., 2002. Evolution of the lectinecomplement pathway and its role in innateimmunity. Nat. Rev. Immunol. 2, 346e353.

Gao, Z., Przeworski, M., Sella, G., 2015. Footprints of ancient-balanced poly-morphisms in genetic variation data from closely related species. Evolution 69,431e446.

Gilpin, W., Feldman, M.W., Aoki, K., 2016. An ecocultural model predicts Neander-thal extinction through competition with modern humans. Proc. Natl. Acad. Sci.113, 2134e2139.

Golovanova, L.V., Doronichev, V.B., Cleghorn, N.E., Koulkova, M.A., Sapelko, T.V.,Shackley, M.S., 2010. Significance of ecological factors in the Middle to UpperPaleolithic transition. Curr. Anthropol. 51, 655e691.

Gravel, S., 2016. When is selection effective? Genetics 203, 451e462.Green, R.E., Krause, J., Briggs, A.W., Maricic, T., Stenzel, U., Kircher, M., Patterson, N.,

Li, H., Zhai, W., Fritz, M.H.Y., Hansen, N.F., Durand, E.Y., Malaspinas, A.S.,Jensen, J.D., Marques-Bonet, T., Alkan, C., Prufer, K., Meyer, M., Burbano, H.A.,Good, J.M., Schultz, R., Aximu-Petri, A., Butthof, A., Hober, B., Hoffner, B.,Siegemund, M., Weihmann, A., Nusbaum, C., Lander, E.S., Russ, C., Novod, N.,Affourtit, J., Egholm, M., Verna, C., Rudan, P., Brajkovic, D., Kucan, Z., Gusic, I.,Doronichev, V.B., Golovanova, L.V., Lalueza-Fox, C., de la Rasilla, M., Fortea, J.,Rosas, A., Schmitz, R.W., Johnson, P.L.F., Eichler, E.E., Falush, D., Birney, E.,Mullikin, J.C., Slatkin, M., Nielsen, R., Kelso, J., Lachmann, M., Reich, D., P€a€abo, S.,2010. A draft sequence of the Neandertal genome. Science 328, 710e722.

Green, R.E., Malaspinas, A., Krause, J., Briggs, A.W., Johnson, P.L.F., Uhler, C.,Meyer, M., Good, J.M., Maricic, T., Stenzel, U., Prüfer, K., Siebauer, M., Burbano, H.a, Ronan, M., Rothberg, J.M., Egholm, M., Rudan, P., Brajkovi�c, D., Ku�can, �Z.,Gu�si�c, I., Wikstr€om, M., Laakkonen, L., Kelso, J., Slatkin, M., P€a€abo, S., 2008.A complete Neandertal mitochondrial genome sequence determined by high-throughput sequencing. Cell 134, 416e426.

Grün, R., Stringer, C., McDermott, F., Nathan, R., Porat, N., Robertson, S., Taylor, L.,Mortimer, G., Eggins, S., McCulloch, M., 2005. U-series and ESR analyses of bonesand teeth relating to the human burials from Skhul. J. Hum. Evol. 49, 316e334.

Gustafsson, K., Widmark, E., Jonsson, A.K., Servenius, B., Sachs, D.H., Larhammar, D.,Rask, L., Peterson, P.A., 1987. Class II genes of the human major histocompati-bility complex: Evolution of the DP region as deduced from nucleotide se-quences of the four genes. J. Biol. Chem. 262, 8778e8786.

Harris, K., Nielsen, R., 2016. The genetic cost of Neanderthal introgression. Genetics203, 881e891.

Henn, B.M., Botigu�e, L.R., Peischl, S., Dupanloup, I., Lipatov, M., Maples, B.K.,Martin, A.R., Musharoff, S., Cann, H., Snyder, M.P., Excoffier, L., Kidd, J.M.,Bustamante, C.D., 2016. Distance from sub-Saharan Africa predicts mutationalload in diverse human genomes. Proc. Natl. Acad. Sci. 113, E440eE449.

Higham, T., Douka, K., Wood, R., Ramsey, C.B., Brock, F., Basell, L., Camps, M.,Arrizabalaga, A., Baena, J., Barroso-Ruíz, C., Bergman, C., Boitard, C., Boscato, P.,Caparr�os, M., Conard, N.J., Draily, C., Froment, A., Galv�an, B., Gambassini, P.,Garcia-Moreno, A., Grimaldi, S., Haesaerts, P., Holt, B., Iriarte-Chiapusso, M.-J.,Jelinek, A., Jord�a Pardo, J.F., Maíllo-Fern�andez, J.-M., Marom, A., Maroto, J.,Men�endez, M., Metz, L., Morin, E., Moroni, A., Negrino, F., Panagopoulou, E.,Peresani, M., Pirson, S., de la Rasilla, M., Riel-Salvatore, J., Ronchitelli, A.,Santamaria, D., Semal, P., Slimak, L., Soler, J., Soler, N., Villaluenga, A., Pinhasi, R.,Jacobi, R., 2014. The timing and spatiotemporal patterning of Neanderthaldisappearance. Nature 512, 306e309.

Horan, R.D., Bulte, E., Shogren, J.F., 2005. How trade saved humanity from biologicalexclusion: an economic theory of Neanderthal extinction. J. Econ. Behav. Organ.58, 1e29.

Houldcroft, C.J., Underdown, S.J., 2016. Neanderthal genomics suggests a Pleistocenetime frame for the first epidemiologic transition. Am. J. Phys. Anthropol. 160,379e388.

Hughes, A.L., Packer, B., Welch, R., Bergen, A.W., Chanock, S.J., Yeager, M., 2003.Widespread purifying selection at polymorphic sites in human protein-codingloci. Proc. Natl. Acad. Sci. 100, 15754e15757.

Juric, I., Aeschbacher, S., Coop, G., 2016. The strength of selection against Nean-derthal introgression. PLOS Genet. 12, e1006340.

Key, F.M., Teixeira, J.C., de Filippo, C., Andr�es, A.M., 2014. Advantageous diversitymaintained by balancing selection in humans. Curr. Opin. Genet. Dev. 29,45e51.

Krings, M., Capelli, C., Tschentscher, F., Geisert, H., Meyer, S., von Haeseler, A.,Grossschmidt, K., Possnert, G., Paunovic, M., P€a€abo, S., 2000. A view of Nean-dertal genetic diversity. Nature Genet. 26, 144e146.

Krings, M., Stone, A., Schmitz, R.W., Krainitzki, H., Stoneking, M., P€a€abo, S., 1997.Neandertal DNA sequences and the origin of modern humans. Cell 90, 19e30.

Kuhlwilm, M., Gronau, I., Hubisz, M.J., de Filippo, C., Prado-Martinez, J., Kircher, M.,Fu, Q., Burbano, H.A., Lalueza-Fox, C., de la Rasilla, M., Rosas, A., Rudan, P.,Brajkovic, D., Kucan, �Z., Gu�sic, I., Marques-Bonet, T., Andr�es, A.M., Viola, B.,P€a€abo, S., Meyer, M., Siepel, A., Castellano, S., 2016. Ancient gene flow from earlymodern humans into Eastern Neanderthals. Nature 530, 429e433.

Lao, O., Lu, T.T., Nothnagel, M., Junge, O., Freitag-Wolf, S., Caliebe, A., Balascakova, M.,Bertranpetit, J., Bindoff, L.A., Comas, D., Holmlund, G., Kouvatsi, A., Macek, M.,Mollet, I., Parson, W., Palo, J., Ploski, R., Sajantila, A., Tagliabracci, A., Gether, U.,Werge, T., Rivadeneira, F., Hofman, A., Uitterlinden, A.G., Gieger, C.,

Wichmann, H.-E., Rüther, A., Schreiber, S., Becker, C., Nürnberg, P., Nelson, M.R.,Krawczak, M., Kayser, M., 2008. Correlation between genetic and geographicstructure in Europe. Curr. Biol. 18, 1241e1248.

Leffler, E.M., Gao, Z., Pfeifer, S., Segurel, L., Auton, A., Venn, O., Bowden, R.,Bontrop, R., Wall, J.D., Sella, G., Donnelly, P., McVean, G., Przeworski, M., 2013.Multiple instances of ancient balancing selection shared between humans andchimpanzees. Science 339, 1578e1582.

Lenz, T.L., Eizaguirre, C., Kalbe, M., Milinski, M., 2013. Evaluating patterns ofconvergent evolution and trans-species polymorphism at MHC immunogenesin two sympatric stickleback species. Evolution 67, 2400e2412.

Loisel, D.A., Rockman, M.V., Wray, G.A., Altmann, J., Alberts, S.C., 2006. Ancientpolymorphism and functional variation in the primate MHC-DQA1 5’ cis-regulatory region. Proc. Natl. Acad. Sci. 103, 16331e16336.

Lowe, J., Barton, N., Blockley, S., Ramsey, C.B., Cullen, V.L., Davies, W., Gamble, C.,Grant, K., Hardiman, M., Housley, R., Lane, C.S., Lee, S., Lewis, M., MacLeod, A.,Menzies, M., Muller, W., Pollard, M., Price, C., Roberts, A.P., Rohling, E.J.,Satow, C., Smith, V.C., Stringer, C.B., Tomlinson, E.L., White, D., Albert, P.,Arienzo, I., Barker, G., Boric, D., Carandente, A., Civetta, L., Ferrier, C., Guadelli, J.-L., Karkanas, P., Koumouzelis, M., Muller, U.C., Orsi, G., Pross, J., Rosi, M., Sha-lamanov-Korobar, L., Sirakov, N., Tzedakis, P.C., 2012. Volcanic ash layers illu-minate the resilience of Neanderthals and early modern humans to naturalhazards. Proc. Natl. Acad. Sci. 109, 13532e13537.

Maricic, T., Gunther, V., Georgiev, O., Gehre, S., Curlin, M., Schreiweis, C.,Naumann, R., Burbano, H.A., Meyer, M., Lalueza-Fox, C., de la Rasilla, M.,Rosas, A., Gajovic, S., Kelso, J., Enard, W., Schaffner, W., P€a€abo, S., 2013. A recentevolutionary change affects a regulatory element in the human FOXP2 gene.Mol. Biol. Evol. 30, 844e852.

Markert, J.A., Grant, P.R., Grant, B.R., Keller, L.F., Coombs, J.L., Petren, K., 2004.Neutral locus heterozygosity, inbreeding, and survival in Darwin's groundfinches (Geospiza fortis and G. scandens). Heredity 92, 306e315.

Marwick, B., 2003. Pleistocene exchange networks as evidence for the evolution oflanguage. Camb. Archaeol. J. 13, 67e81.

Meyer, M., Kircher, M., Gansauge, M.-T., Li, H., Racimo, F., Mallick, S., Schraiber, J.G.,Jay, F., Prufer, K., de Filippo, C., Sudmant, P.H., Alkan, C., Fu, Q., Do, R.,Rohland, N., Tandon, A., Siebauer, M., Green, R.E., Bryc, K., Briggs, A.W.,Stenzel, U., Dabney, J., Shendure, J., Kitzman, J., Hammer, M.F., Shunkov, M.V.,Derevianko, A.P., Patterson, N., Andres, A.M., Eichler, E.E., Slatkin, M., Reich, D.,Kelso, J., P€a€abo, S., 2012. A high-coverage genome sequence from an archaicDenisovan individual. Science 338, 222e226.

Morelli, G., Song, Y., Mazzoni, C.J., Eppinger, M., Roumagnac, P., Wagner, D.M.,Feldkamp, M., Kusecek, B., Vogler, A.J., Li, Y., Cui, Y., Thomson, N.R., Jombart, T.,Leblois, R., Lichtner, P., Rahalison, L., Petersen, J.M., Balloux, F., Keim, P., Wirth, T.,Ravel, J., Yang, R., Carniel, E., Achtman, M., 2010. Yersinia pestis genomesequencing identifies patterns of global phylogenetic diversity. Nature Genet.42, 1140e1143.

Nei, M., Li, W.H., 1979. Mathematical model for studying genetic variation in termsof restriction endonucleases. Proc. Natl. Acad. Sci. 76, 5269e5273.

Netea, M.G., Joosten, L.A.B., 2016. TLRs of our fathers. Immunity 44, 218e220.Nielsen, R., Akey, J.M., Jakobsson, M., Pritchard, J.K., Tishkoff, S., Willerslev, E., 2017.

Tracing the peopling of the world through genomics. Nature 541, 302e310.Parks, M., Subramanian, S., Baroni, C., Salvatore, M.C., Zhang, G., Millar, C.D.,

Lambert, D.M., 2015. Ancient population genomics and the study of evolution.Phil. Trans. R. Soc. B 370, 20130381.

Perry, R.D., Fetherston, J.D., 1997. Yersinia pestis-etiologic agent of plague. Clin.Microbiol. Rev. 10, 35e66.

Ponce de Leon, M.S., Golovanova, L., Doronichev, V., Romanova, G., Akazawa, T.,Kondo, O., Ishida, H., Zollikofer, C.P.E., 2008. Neanderthal brain size at birthprovides insights into the evolution of human life history. Proc. Natl. Acad. Sci.105, 13764e13768.

Prado-Martinez, J., Sudmant, P.H., Kidd, J.M., Li, H., Kelley, J.L., Lorente-Galdos, B.,Veeramah, K.R., Woerner, A.E., O'Connor, T.D., Santpere, G., Cagan, A., Theunert, C.,Casals, F., Laayouni, H., Munch, K., Hobolth, A., Halager, A.E., Malig, M., Hernan-dez-Rodriguez, J., Hernando-Herraez, I., Prüfer, K., Pybus, M., Johnstone, L.,Lachmann, M., Alkan, C., Twigg, D., Petit, N., Baker, C., Hormozdiari, F., Fernandez-Callejo, M., Dabad, M., Wilson, M.L., Stevison, L., Camprubí, C., Carvalho, T., Ruiz-Herrera, A., Vives, L., Mele, M., Abello, T., Kondova, I., Bontrop, R.E., Pusey, A.,Lankester, F., Kiyang, J.A., Bergl, R.A., Lonsdorf, E., Myers, S., Ventura, M.,Gagneux, P., Comas, D., Siegismund, H., Blanc, J., Agueda-Calpena, L., Gut, M.,Fulton, L., Tishkoff, S.A., Mullikin, J.C., Wilson, R.K., Gut, I.G., Gonder, M.K.,Ryder, O.A., Hahn, B.H., Navarro, A., Akey, J.M., Bertranpetit, J., Reich, D.,Mailund, T., Schierup, M.H., Hvilsom, C., Andr�es, A.M., Wall, J.D., Bustamante, C.D.,Hammer, M.F., Eichler, E.E., Marques-Bonet, T., 2013. Great ape genetic diversityand population history. Nature 499, 471e475.

Prüfer, K., Racimo, F., Patterson, N., Jay, F., Sankararaman, S., Sawyer, S., Heinze, A.,Renaud, G., Sudmant, P.H., de Filippo, C., Li, H., Mallick, S., Dannemann, M.,Fu, Q., Kircher, M., Kuhlwilm, M., Lachmann, M., Meyer, M., Ongyerth, M.,Siebauer, M., Theunert, C., Tandon, A., Moorjani, P., Pickrell, J., Mullikin, J.C.,Vohr, S.H., Green, R.E., Hellmann, I., Johnson, P.L.F., Blanche, H., Cann, H.,Kitzman, J.O., Shendure, J., Eichler, E.E., Lein, E.S., Bakken, T.E., Golovanova, L.V.,Doronichev, V.B., Shunkov, M.V., Derevianko, A.P., Viola, B., Slatkin, M., Reich, D.,Kelso, J., P€a€abo, S., 2014. The complete genome sequence of a Neanderthal fromthe Altai Mountains. Nature 505, 43e49.

Qutob, N., Balloux, F., Raj, T., Liu, H., Marion de Proc�e, S., Trowsdale, J., Manica, A.,2012. Signatures of historical demography and pathogen richness on MHC classI genes. Immunogenetics 64, 165e175.

A.P. Sullivan et al. / Journal of Human Evolution 108 (2017) 62e71 71

Racimo, F., Marnetto, D., Huerta-S�anchez, E., 2017. Signatures of archaic adaptiveintrogression in present-day human populations. Mol. Biol. Evol. 34, 296e317.

R Core Team, 2015. R: A language and environment for statistical computing[WWW Document]. R Foundation for Statistical Computing, Vienna, Austria.http://www.r-project.org/.

Raichlen, D.A., Armstrong, H., Lieberman, D.E., 2011. Calcaneus length determinesrunning economy: Implications for endurance running performance in modernhumans and Neandertals. J. Hum. Evol. 60, 299e308.

Ramenofsky, A., Wilbur, A.K., Stone, A.C., 2003. Native American disease history:past, present and future directions. World Archaeol. 35, 241e257.

Rasmussen, M.D., Hubisz, M.J., Gronau, I., Siepel, A., 2014. Genome-wide inferenceof ancestral recombination graphs. PLOS Genet. 10, e1004342.

Sams, A.J., Dumaine, A., N�ed�elec, Y., Yotova, V., Alfieri, C., Tanner, J.E., Messer, P.W.,Barreiro, L.B., 2016. Adaptively introgressed Neandertal haplotype at the OASlocus functionally impacts innate immune responses in humans. Genome Biol.17, 246.

Sandgathe, D.M., Dibble, H.L., Goldberg, P., McPherron, S.P., Turq, A., Niven, L.,Hodgkins, J., 2011. On the role of fire in Neandertal adaptations in WesternEurope: Evidence from Peche de l’Az�e IV and Roc de Marsal, France. PaleoAn-thropology 216e242.

Sankararaman, S., Patterson, N., Li, H., P€a€abo, S., Reich, D., 2012. The date of inter-breeding between Neandertals and modern humans. PLoS Genet. 8, e1002947.

Seguin-Orlando, A., Korneliussen, T.S., Sikora, M., Malaspinas, A.-S., Manica, A.,Moltke, I., Albrechtsen, A., Ko, A., Margaryan, A., Moiseyev, V., Goebel, T.,Westaway, M., Lambert, D., Khartanovich, V., Wall, J.D., Nigst, P.R., Foley, R.A.,Lahr, M.M., Nielsen, R., Orlando, L., Willerslev, E., 2014. Genomic structure inEuropeans dating back at least 36,200 years. Science 346, 1113e1118.

Shipman, P., 2015. How do you kill 86 mammoths? Taphonomic investigations ofmammoth megasites. Quatern. Int. 359e360, 38e46.

Simonti, C.N., Vernot, B., Bastarache, L., Bottinger, E., Carrell, D.S., Chisholm, R.L.,Crosslin, D.R., Hebbring, S.J., Jarvik, G.P., Kullo, I.J., Li, R., Pathak, J., Ritchie, M.D.,Roden, D.M., Verma, S.S., Tromp, G., Prato, J.D., Bush, W.S., Akey, J.M., Denny, J.C.,Capra, J.A., 2016. The phenotypic legacy of admixture between modern humansand Neandertals. Science 351, 737e741.

Thornton, R., 1997. Aboriginal North American population and rates of decline, ca.A.D. 1500e1900. Curr. Anthropol. 38, 310e315.

Trinkaus, E., 1984. Neandertal pubic morphology and gestation length. Curr.Anthropol. 25, 509e514.

Tzedakis, P.C., Hughen, K.A., Cacho, I., Harvati, K., 2007. Placing late Neanderthals ina climatic context. Nature 449, 206e208.

Valet, J.-P., Valladas, H., 2010. The Laschamp-Mono lake geomagnetic events and theextinction of Neanderthal: a causal link or a coincidence? Quaternary Sci. Rev.29, 3887e3893.

Vernot, B., Akey, J.M., 2015. Complex history of admixture between modern humansand Neandertals. Am. J. Hum. Genet. 96, 448e453.

Wolff, H., Greenwood, A.D., 2010. Did viral disease of humans wipe out the Nean-dertals? Med. Hypotheses 75, 99e105.

Wood, R.E., Higham, T.F.G., de Torres, T., Tisnerat-Laborde, N., Valladas, H., Ortiz, J.E.,Lalueza-Fox, C., Sanchez-Moral, S., Canaveras, J.C., Rosas, A., Santamaria, D., de laRasilla, M., 2013. A new date for the Neanderthals from El Sidr�on Cave (Asturias,Northern Spain). Archaeometry 55, 148e158.

Zhao, Z., Fu, Y.X., Hewett-Emmett, D., Boerwinkle, E., 2003. Investigating singlenucleotide polymorphism (SNP) density in the human genome and its impli-cations for molecular evolution. Gene 312, 207e213.