Gene 530 (2013) 83–94

Contents lists available at ScienceDirect

Gene
journal homepage: www.elsevier.com/locate/gene

Neanderthal and Denisova genetic affinities with contemporary humans:

Introgression versus common ancestral polymorphisms
Robert K. Lowery a,b,c,⁎, Gabriel Uribe d,1, Eric B. Jimenez e,1, Mark A. Weiss f, Kristian J. Herrera g,
Maria Regueiro a, Rene J. Herrera a
a
Department of Molecular and Human Genetics, College of Medicine, Florida International University, Miami, FL 33199, USA
b
Department of Biological Sciences, College of Arts and Sciences, Florida International University, Miami, FL 33199, USA
c
Department of Biological Sciences, College of Arts and Sciences, Indian River State College, Fort Pierce, FL 34981, USA
d
TERRA Environmental Research Institute, Miami, FL 33173, USA
e
Department of Molecular and Cellular Biology, Wake Forest University Graduate School of Arts and Sciences, Medical Center Boulevard, Winston Salem, NC 27157, USA
f
School of Computing and Information Sciences, Florida International University, Miami, FL 33199, USA
g
Department of Molecular and Cell Biology, Harvard University, Cambridge, MA 02138, USA

a r t i c l e i n f o a b s t r a c t

Article history: Analyses of the genetic relationships among modern humans, Neanderthals and Denisovans have suggested that
Accepted 11 June 2013 1–4% of the non-Sub-Saharan African gene pool may be Neanderthal derived, while 6–8% of the Melanesian gene
Available online 19 July 2013 pool may be the product of admixture between the Denisovans and the direct ancestors of Melanesians. In the
present study, we analyzed single nucleotide polymorphism (SNP) diversity among a worldwide collection of
Keywords:
contemporary human populations with respect to the genetic constitution of these two archaic hominins and
Neanderthal
Neandertal
Pan troglodytes (chimpanzee). We partitioned SNPs into subsets, including those that are derived in both archaic
Archaic lineages, those that are ancestral in both archaic lineages and those that are only derived in one archaic lineage.
Denisova By doing this, we have conducted separate examinations of subsets of mutations with higher probabilities of di-
Phylogenetics vergent phylogenetic origins. While previous investigations have excluded SNPs from common ancestors in prin-
Human evolution cipal component analyses, we included common ancestral SNPs in our analyses to visualize the relative
placement of the Neanderthal and Denisova among human populations. To assess the genetic similarities
among the various hominin lineages, we performed genetic structure analyses to provide a comparison of genetic
patterns found within contemporary human genomes that may have archaic or common ancestral roots. Our re-
sults indicate that 3.6% of the Neanderthal genome is shared with roughly 65.4% of the average European gene
pool, which clinally diminishes with distance from Europe. Our results suggest that Neanderthal genetic associ-
ations with contemporary non-Sub-Saharan African populations, as well as the genetic affinities observed be-
tween Denisovans and Melanesians most likely result from the retention of ancient mutations in these
populations.
© 2013 Elsevier B.V. All rights reserved.

1. Introduction

The release of the draft assemblies of the Neanderthal (Green et al.,

Abbreviations: SNP, single nucleotide polymorphism; SNPs, single nucleotide polymor-
phisms; DNA, deoxyribonucleic acid; NdDa, Neanderthal derived and Denisova ancestral;
NaDd, Neanderthal ancestral and Denisova derived; NdDd, Neanderthal derived and
Denisova derived; NaDa, Neanderthal ancestral and Denisova ancestral; hg18, human
genome (18th assembly by the University of California, Santa Cruz); UCSC, University of
California, Santa Cruz; PCA, principal component analysis; PCAs, principal component
analyses; PC1, 1st principal component; PC2, 2nd principal component; PC3, 3rd principal
component; kya, thousand years ago.
⁎ Corresponding author at: Department of Biological Sciences, College of Arts and Sciences,
Indian River State College, 3209 Virginia Ave., N 207, Fort Pierce, FL 34981, USA. Tel.: +1 772
462 7122; fax: +1 772 462 7447.
E-mail addresses: rlowe001@fiu.edu, rlowery@irsc.edu (R.K. Lowery),
gaurgo2007@gmail.com (G. Uribe), eric@ericjimenez.com (E.B. Jimenez), Weiss@fiu.edu
(M.A. Weiss), kherrera01@fas.harvard.edu (K.J. Herrera), mmregueiro@gmail.com
(M. Regueiro), herrerar@fiu.edu (R.J. Herrera).
1
These two authors contributed equally to this manuscript.
2010) and Denisova genomes (Reich et al., 2010) was accompanied by
statistical analyses that sought to assess both the occurrence and extent
of introgression from these archaic species into contemporary human
populations. Most notably, these analyses employed a statistic tool
called D-statistics (Green et al., 2010) which was designed to determine
the genetic relatedness of two populations relative to a third outlier
group and an ancestral control (Pan troglodytes). In the first of these
studies (Green et al., 2010), D-statistics were compared among contem-
porary human populations using the frequencies of derived alleles in
single nucleotide polymorphisms (SNPs) (i.e., the alleles not present
in P. troglodytes) also present in the archaic genome of interest. These
sets were defined under the assumption that in SNPs where the derived
state is present in both archaic hominid and contemporary human

0378-1119/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.gene.2013.06.005
84 R.K. Lowery et al. / Gene 530 (2013) 83–94
populations, the derived alleles are most likely to include products of
introgression (Green et al., 2010; Reich et al., 2010).
These analyses suggested that non-Sub-Saharan Africans contain
1–4% higher Neanderthal derived allelic frequencies than Sub-
Saharan-Africans. The potential explanation for these observations
that was provided by Green et al. (2010) is that modern human popu-
lation(s) admixed with Neanderthals near the time of or shortly
after migrating out of Africa. In this scenario, all descendants of this
ancient modern human population inherited roughly equal levels of
Neanderthal DNA. The primary assumption of this model is that the
analyzed set of mutations which establish the D-statistic based differ-
ences among Sub-Saharan Africans and non-Sub-Saharan Africans in a
hypothetical Neanderthal gene flow event(s) involved SNPs that origi-
nated in the Neanderthal lineage. However, simply because a contempo-
rary human SNP allele is not seen within the chimpanzee genome but is
found within the Neanderthal lineage does not mean that the mutation
necessarily first occurred within Neanderthals. Such mutant alleles
could have originated any time between the divergence of the chimpan-
zee and hominin lineages, 4–7 million years ago (Langergraber et al.,
2012), and the split between Neanderthals and modern humans
(400,000 to 800,000 years ago) (Green et al., 2010). Such derived alleles
may have survived (without reaching fixation) to enter the gene pool of
a recent common ancestor of Neanderthals and contemporary humans.
Thus, as noted by Green et al. (2010), it is possible that many Neander-
thal derived alleles arose before Neanderthals had evolved. In other
words, the possession of these mutations in both archaic groups and
contemporary humans may represent common ancestral polymor-
phisms. While facing similar caveats, Reich et al. (2010) in their analyses
of the Denisova genome, found evidence supporting 6–8% higher
frequencies of Denisova derived alleles in the Melanesian population
relative to other contemporary groups, and suggested that this is the
result of admixture between Denisovans and the Melanesians ancestors
as they migrated from Africa along the “Southern Coastal Route” of
Southern Asia (Cann, 2001; Cavalli-Sforza et al., 1994; Kennedy et al.,
1987; Kong et al., 2006; Lahr and Foley, 1994; Maca-Meyer et al.,
2001; Macaulay et al., 2005; Palanichamy et al., 2004; Quintana-Murci
et al., 1999; Stringer, 2000; Sun et al., 2006; Tanaka et al., 2004). Reich
et al. (2010) also streamlined the set of SNPs employed in their
Neanderthal admixture tests to include only SNPs that exhibited the
derived state in the Neanderthal and the ancestral state in the
Denisovan. Based on these analyses, Reich et al. (2010) generated
D-statistic-based results similar to those of Green et al. (2010) indicating
1–4% greater Neanderthal relatedness in non Sub-Saharan-Africans.
As stated by the above mentioned investigators, an alternative to
introgression as an explanation for the observed affinities between
archaic hominins like Neanderthals or Denisovans and contemporary
humans is the existence of ancestral common polymorphisms that
managed to infiltrate multiple lineages. In terms of the differences in
the level of genetic similarities to Neanderthal between non Sub-
Saharan Africans and Sub-Saharan Africans, it is possible that regional
differentiation in Africa within the common ancestral population of all
modern humans occurred at a time early enough for some human pop-
ulations to be genetically closer to Neanderthals or Denisovans than
other populations. This could have been facilitated by the existence of
ancient genetic sub-population structure within modern humans prior
to the spread of modern humans out of Africa (Durand et al., 2011;
Slatkin and Pollack, 2008). In other words, the humans that migrated
out of Africa may have represented a subpopulation with greater affin-
ities to Neanderthals or Denisovans. Other forces like genetic drift, foun-
der effects and/or bottle neck events could have been responsible for
the split of ancestral modern humans into subpopulations and/or gener-
ated contemporary human populations with greater proportions of ar-
chaic hominid DNA during the peopling of the world.
Investigators have tested these hypotheses (introgression versus
common ancestral polymorphism followed by some form of genetic
drift) by modeling population dynamics and applying new statistical
methods and have come to divergent conclusions. For example, Durand
et al. (2011) concludes that either of the two hypotheses is possible,
as D-statistic analysis is susceptible to the influence of ancestral genetic
subpopulation structure. Yang et al. (2012) later reported that only the
admixture-based model could account for the frequency spectrums of
Neanderthal derived alleles seen in contemporary human populations.
However, Yang et al. (2012) employed a uniform coalescence time of
112,000 years for all contemporary human populations, which prevents
the analysis of a model in which the formation of the ancestral subpopu-
lation structure predates the emergence of modern humans from Africa.
Furthermore, recent simulations of the different demographic models
that account for a potentially older divergence time indicated that the
D-statistic differences observed between non Sub-Saharan Africans and
Sub-Saharan Africans are, in fact, compatible with a model that does not
incur admixture (Eriksson and Manica, 2012).
In the present study, we examine the possible origins of genetic
similarities between groups of contemporary humans (n = 827 indi-
viduals belonging to 11 metapopulations) and archaic hominins by ex-
ploring their patterns of genetic relatedness. We subdivided SNP-based
genetic diversity into four SNP subsets that are likely to have originated
at different time points in recent human evolution. These four subsets
include SNPs that are derived in the Neanderthal and ancestral in the
Denisovan (NdDa), those that are ancestral in Neanderthals but derived
in the Denisovan (NaDd), those that are derived in both Neanderthals
and the Denisovan (NdDd) and those that are ancestral in both Nean-
derthals and the Denisovan (NaDa). By so doing, we have attempted
to examine sets of SNPs most likely to stem from the common ancestors
with modern humans. We also examined the genetic affinities between
the archaic hominins and contemporary human wherein the ancestral
component is not edited out. D-statistics, principal component analyses
and genetic structure-based observations are utilized to assess whether
the observed genetic similarities are more consistent with admixture or
common ancestral polymorphisms. In this study, we utilize genetic
Structure analysis for the first time to investigate the specific compo-
nents of the human genome which are shared with Neanderthals and
the Denisova.
2. Materials and methods
2.1. Neanderthal and Denisovan ancestral and derived alleles in
contemporary human populations
To address the question of whether genetic similarities between ar-
chaic hominins and contemporary humans resulted from introgression
or common ancestral polymorphisms retained in current human popu-
lations, we procured roughly five million SNPs from the selective sweep
analysis conducted in the Green et al. (2010) draft assembly of the
Neanderthal genome. The genotypes for the Neanderthal and the Refer-
ence Human at these loci were provided (Green et al., 2010) as ancestral
or derived based on comparison to P. troglodytes. By contrasting of the
derived or ancestral status of the Reference Genome genotype at each
SNP, the Neanderthal genotype was identified as the same or different
to the Reference Human. The approximately five million human SNPs
for which the Neanderthal genotype was acquired (Green et al., 2010)
were employed to create a consensus dataset of SNPs from 827
human samples representing a comprehensive group of worldwide
contemporary populations (Behar et al., 2010; Jakobsson et al., 2008).
A list of populations included in our analyses and corresponding source
literature is provided in Table 1. Behar et al. (2010) utilized only SNPs
that were unambiguous with respect to strand orientation (i.e., no
SNPs with complementary alleles). For consistency, all strand ambigui-
ties were removed from our analyses during the creation of a consensus
of the SNPs from Behar et al. (2010), Jakobsson et al. (2008) and those
from the Neanderthal selective sweep analyses. The chromosomal posi-
tions for these consensus SNPs were then employed to find the geno-
type of the Denisova Genome sequences, which were mapped against
 R.K. Lowery et al. / Gene 530 (2013) 83–94 85

Table 1 the entire set of 37,758 SNPs as well as the four subsets of SNPs
List of modern regional metapopulations used in this study. (NaDa, NdDa, NaDd, and NdDd) individually. The resultant PCA plots
Abbreviation n Region Populations included are presented in Figs. 1 through 5, respectively. PCA is a specific type
a 123 Sub-Saharan Africa Yoruba, Mandenka, San, Bantu,
of correspondence analysis wherein the primary difference among
Biaka Pygmy, Mbuti Pygmy categories, based upon certain factors (in this case SNP genotypes),
b 41 Northern Africa Ethiopians, Egypt, Morocco can be used to plot individuals along a single axis representative of
c 68 Caucasus Georgia, Armenians, Lezgins, Adygei that primary variance (PC1). With the same data set, subsequent, lesser
e 124 Europe Lithuanian, Belorussian, Romanian,
differences among individuals may be plotted to construct additional
Cypriot, Hungarian, Basque, Russian,
Spanish, Chuvash axes (PC2, PC3, etc…), for visual comparison. In parallel, the genetic
m 33 Melanesian Papuan, Bouganville structures of the contemporary human populations were examined in
n 31 Amerindian Pima, Piapoco, Curripaco, Mayan relation to the two archaic hominins and the chimpanzee using the pro-
s 67 South Central Asia Paniya, Kannadi, Sakilli, Kalash, Uygur, gram STRUCTURE v. 2.3.4 (http://pritch.bsd.uchicago.edu/structure.
Barusho, Balochi
html). Genetic structure diagrams are presented in Figs. 6 through 10,
r 35 South West Asia Iranian, Uzbekistan
d 30 South East Asia Yizu, Cambodian, Lahu, Malayan respectively. The STRUCTURE program requires the input of a value
t 34 North East Asia Yakut, Mongolian, Daur for the parameter k which represents the number of putative ancestral
z 241 Near-East Jordan, Samaritan, Syrian, Druze, populations contributing to the dataset and the number of genetic pat-
Bedouin, Mozabite, Palestinian,
terns or components which the program will search for among all indi-
Turkey, Lebanon, Saudi, Yemen
viduals of the dataset. In the Structure analyses, all individuals were
treated as equal participants in any potential gene-flow/admixture
Build 36.1 of the human genome (hg18) (Reich et al., 2010) and provided event as opposed to being restricted in gene-flow to their respective
by the Table Browser of the UCSC Genome Browser. In cases where the populations. The “admixture” option of the STRUCTURE program was
multiple Neanderthal or Denisovan sequences at a given SNP locus pro- selected because 1) restriction of sets of humans to defined populations
vided conflicting or ambiguous genotypes of either the Neanderthal or introduces bias in the assessment of similarities among all individuals
the Denisovan, these SNPs were removed from further analyses. This gen- (and taxa) tested and 2) putative admixture events could have preced-
erated a set of 96,257 SNPs for which the Neanderthal, Denisovan and the ed the partitioning of modern humans into current-day populations. To
collection of 827 human genotypes was known. These 96,257 SNPs were investigate a large enough set of genetic similarities required to evaluate
reduced to a final set of 37,758 by including only those SNPs kept by any archaic relationship to contemporary humans, a wide range of k
Behar et al. (2010) after additional stringency tests were implemented values (2 through 15) were examined. We report the structure results
to adjust for ascertainment bias and reduce the influence of linkage dis- at k = 10 since higher k values provided no additional partitioning of
equilibrium upon evaluation of genetic structure and principal compo- populations and k values below 10 yield less definition of populations.
nent analysis. No SNPs from the X chromosome, Y chromosome and
mitochondrial genome were examined in this study. 3. Results and discussion
The SNP genotypes of the 827 modern humans, the chimpanzee, the
Neanderthal and the Denisovan genomes were divided into four subsets 3.1. Evidence for Neanderthal gene flow
as follows: 1) SNPs that are derived in the Neanderthal and ancestral in
the Denisovan (NdDa SNPs; n = 2776), 2) those that are ancestral in From the distribution of component 7 in the structure diagram gen-
Neanderthals but derived in the Denisovan (NaDd SNPs; n = 2322), erated from all 37,758 SNPs, we observe that 3.6% (Fig. 6, in aqua color)
3) those that are derived in both Neanderthals and the Denisovan of the Neanderthal profile is shared primarily with Europeans (average
(NdDd SNPs; n = 4746), and 4) those that are ancestral in both 65.4%), with lower levels in the Caucasia (42.5%), followed by the South
Neanderthals and the Denisovan (NaDa SNPs; n = 27,914). Each SNP West Asia (29.7%) and the Near East (23.5%). This gradient is compatible
subset is expected to reflect a different range of time points in the evo- with introgression or common ancestral polymorphisms shared between
lution of the three hominin lineages, with the expected time points the ancestors of the Modern Europeans and Neanderthals, followed
dependent upon which model of evolution is assumed. In particular, by demic diffusion into other modern human populations eastward.
we propose that the NdDa mutations most likely originated in either The existence of a cline of genetic similarity to Neanderthals that is
the Neanderthal lineage, if the derived allele entered the human ge- highest in Europe is further supported by the STRUCTURE analysis of
nome from Neanderthal admixture, or was present in the common the NdDa subset (Fig. 8) which indicates that the largest Neanderthal
ancestral gene pool of Neanderthals and contemporary humans. Simi- component (7 in light blue; 47.8%) is shared largely with Europeans
larly, NaDd mutations in contemporary humans most likely stem from (73.1%). The observed geographical gradient of affinities between
either human admixture with the Denisova lineage or with a common Neanderthal and contemporary human populations also may have been
ancestor to modern humans. NdDd SNPs, meanwhile, may have mutat- impacted by longer temporal overlap of the two groups in Europe
ed in either the common ancestral gene pool of both Neanderthals leading to higher admixture and a greater degree of genetic similarities.
and Denisova or their common ancestor with modern humans. Alterna- The NdDa Structure graph (Fig. 8) also illustrates that other non Sub-
tively, it should be noted that it is also possible that a NdDd mutation Saharan African populations have notable, albeit lower, affinities with
originated in any one of these three lineages and subsequently entered a lesser Neanderthal component (6 in light brown; 13.1%), particularly
the other two lineages through interbreeding. Such a scenario is the in Amerindians (94.5%), North East Asians (72.3%) and South East Asians
least parsimonious since it would require at least two separate inter- (64.8%). Melanesians carry a component (3 in yellow; 93.7%) which is
breeding events to have occurred. The derived state in the NaDa set of present at 7.5% in the Neanderthal.
SNPs may have mutated in the human lineage at any point up to nearly
the present day. This set of SNPs should establish the common ancestral 3.2. The origin of the NdDd subset of SNPs
state with modern humans since the derived state is only found among
contemporary humans. There are two plausible scenarios for the genesis of the SNPs that are
derived in Neanderthals, Denisova and contemporary humans: 1) they
2.2. Principal component and structure analyses originated in common ancestors or 2) they may stem from gene flow
among the three groups. Of course, a combination of these two scenar-
Principal component analyses (PCAs) were performed using the ios could contribute to the origin of the NdDd subset. The D-statistic re-
EIGENSOFT program (Price et al., 2006; Patterson et al., 2006) utilizing sults demonstrate a higher presence of NdDd alleles in Sub-Saharan
86 R.K. Lowery et al. / Gene 530 (2013) 83–94

Fig. 1. Principal component (PC) analysis of 827 contemporary humans, Neanderthals, the Denisova and chimpanzee using the set of all 37,758 SNPs. PC1 and PC2 comprise 58.2% and
26.17% of variance, respectively.

Africans relative to all Eurasians and North Africans. In a Eurasian ad-
mixture scenario, Sub-Saharan Africans are expected to exhibit fewer
NdDd alleles than Eurasians and North Africans. Based on this, the levels
of NdDd alleles in Sub-Saharan Africans strongly support an ancient
common hominin origin rather than recent tri-lineage archaic admixture
events. Moreover, Sub-Saharan Africans share 3 of the 5 NdDd Structure
components comprising roughly 30% of the NdDd profiles (components
1, 6, and 10 in Fig. 10), which suggests common ancestral origins
rather than tri-partite introgression events for the NdDd SNPs. Addi-
tional support for an ancestral common polymorphism scenario is

Fig. 2. Principal component (PC) analysis of 827 contemporary humans, Neanderthals and the Denisova using the ancestral state (identical to the chimpanzee) of 27,914 SNPs (NaDa). PC1
and PC2 comprise 57.5% and 26.3% of variance, respectively. Full circle represent more than one group.
 R.K. Lowery et al. / Gene 530 (2013) 83–94 87

Fig. 3. Principal component (PC) analysis of 827 contemporary humans, Neanderthals and the Denisova using 2776 SNPs that possess the derived state (different from the chimpanzee) in
Neanderthals and the ancestral state in the Denisova (NdDa). PC1 and PC2 comprise 60.5% and 24.9% of variance, respectively. Full circle represent more than one group.

evident in the location of Neanderthals and the Denisova in the NdDd
plot along PC1 (Fig. 5) in an ancestral position with Sub-Saharan
Africans. This result is surprising given that these genotypes consists
only of derived alleles. Interestingly, the NdDd SNPs exhibit compara-
ble D-statistic-based presence in Melanesians and Sub-Saharan Africans.
Both Melanesians and Sub-Saharan Africans generate higher NdDd
based D-statistics than all other worldwide populations, ranging
from 0.7% greater than North Africans to 1.5% greater than North
East Asians (Table 2; Fig. 11). Furthermore, if tri-lineage admixture
events were a prominent source for NdDd SNPs, we would expect a

Fig. 4. Principal component (PC) analysis of 827 contemporary humans, Neanderthals and the Denisova using 2322 SNPs that possess the derived state (different from the chimpanzee) in
the Denisova and the ancestral state in Neanderthals (NaDd). PC1 and PC2 comprise 57.8% and 26.8% of variance, respectively. Full circle represent more than one group.
88 R.K. Lowery et al. / Gene 530 (2013) 83–94

Fig. 5. Principal component (PC) analysis of 827 contemporary humans, the Neanderthals and the Denisova using 4746 SNPs that possess the derived state (different from the chimpanzee)
in both the Neanderthals and the Denisova (NdDd). PC1 and PC2 comprise 61.7% and 24.7% of variance, respectively. Full circle represent more than one group.

greater degree of genetic sharing between Europeans and Melane-
sians; however, in all of our Structure analyses, no such sharing is
observed.
3.3. Ancient genetic components in Melanesians revealed by SNP subsets
In the analyses of the NdDd SNP subset, we observe both PCA and D-
statistic based similarities between Melanesians and Sub-Saharan
Africans. In the NdDd PCA diagrams (Fig. 5; Supplementary Fig. 5),
Neanderthals and the Denisova plot proximally to both the Sub-
Saharan populations along PC1 and the Melanesian population along
PC3. Further, both Melanesians and Sub-Saharans have elevated NdDd
based D-statistics as opposed to the other worldwide populations.
As seen in prior research (Li et al., 2008) and all of our PCAs, the Sub-
Saharan populations clearly display the greatest proportion of ancestral
human DNA as seen by their proximity to Pan. This is further supported
by the high frequencies of the Pan genome in the two pygmy populations
(component 2 in red, Fig. 7) since they possess one of the oldest gene
pools in Africa. Thus, it is likely that the relatively elevated levels of
NdDd alleles in Sub-Saharans may be due to the presence of common

Fig. 6. STRUCTURE analysis of 827 contemporary humans, Neanderthals and the Denisova using the set of all 37,758 SNPs (A). The number of putative ancestral populations (k) is 10. A
closer view of the components of the Neanderthal, Denisova and chimpanzee profiles is shown to the left.
 R.K. Lowery et al. / Gene 530 (2013) 83–94 89

Fig. 7. STRUCTURE analysis of 827 contemporary humans, Neanderthals and the Denisova using the ancestral state (identical to the chimpanzee) of 27,914 SNPs (NaDa). The number of
putative ancestral populations (k) is 10.

ancestral mutations, which would be consistent with our postulate that
NdDd SNPs represent primarily common ancestral polymorphisms. Fur-
thermore, alternative explanations for the presence of NdDd alleles in
these groups and contemporary humans require more complex inter-
breeding scenarios involving tri-partite gene flow.
Along the PC3 of the NdDd plot (Supplementary Fig. 5), the proxim-
ity of the Denisova and Neanderthals to the Melanesians suggests that
the derived state in Neanderthal/Denisova, signals a closer relationship
with the Melanesians than the ancestral state (Pan). The fact that
this Denisova/Melanesian kinship is demonstrated by a SNP subset
which appears to derive from a common ancestor (NdDd) provides fur-
ther support for the idea that the Denisova are associating closer with
the Melanesians simply because the Melanesians share more derived
alleles from a common ancestor than do other contemporary human
populations
In addition to the support for an excess of common ancestral poly-
morphisms in Melanesians relative to other Eurasians from the D-
statistics results, the NaDd PCA also partitions the Denisovan in close
proximity to Sub-Saharan Africans along PC1 (Fig. 4). This is in stark
contrast to the far less proximal location of Sub-Saharans relative to Ne-
anderthal in the NdDa PCA (Fig. 3). This disparity suggests that many of
the derived mutations present in the Denisova are also found in the Sub-
Saharan African populations, which would dictate an early, common an-
cestral origin for these polymorphisms.
Perhaps due to genetic drift via bottle neck events and/or founder
effects followed by geographic isolation, Melanesians became primarily

Fig. 8. STRUCTURE analysis of 827 contemporary humans, Neanderthals and the Denisova using 2776 SNPs that possess the derived state (different from the chimpanzee) in Neanderthals
and the ancestral state in the Denisova (NdDa). The number of putative ancestral populations (k) is 10.
90 R.K. Lowery et al. / Gene 530 (2013) 83–94

Fig. 9. STRUCTURE analysis of 827 contemporary humans, Neanderthals and the Denisova using 2322 SNPs that possess the derived state (different from the chimpanzee) in the Denisova
and the ancestral state in Neanderthals (NaDd). The number of putative ancestral populations (k) is 10.

composed of one genetic component as seen in each of the five Struc-
ture analyses. For example, component 7 (aqua color in Fig. 10), present
at 5.2% in the Neanderthals/Denisova profile in the NdDd Structure plot,
is seen at an average of 91.9% in Melanesians, while, in the NdDa SNP
subset, component 3 (yellow in Fig. 8) is seen at 93.7% in Melanesians.
The simplest explanation for Melanesians exhibiting primarily a
Neanderthal and Denisova component in the NdDa and NdDd Structure
analyses is that these derived alleles originated primarily in the com-
mon ancestor of Neanderthals, Denisovan and contemporary humans,
and is at higher levels in Melanesians than in other worldwide popula-
tions possibly as a result of genetic drift.
Recent studies reported that East Asians have 24% (Meyer et al.,
2012) and 40% (Wall et al., 2013) greater Neanderthal relatedness
than do Europeans, while others have indicated that East Asians
have higher affinity to Denisovans than do Europeans (Skoglund and
Jakobsson, 2011). Considering our results, the most likely explanation
for a higher Denisovan affinity to East Asians is the retention of
higher levels of common ancestral DNA than Europeans. Further,
if the putative Neanderthal genetic affinity to contemporary human
populations does stem from common ancestral DNA, then East Asians
would appear to have inherited proportionally greater SNPs in common
with Neanderthals than Europeans. In this regard, we do see evidence
for a component (4) in the overall set of SNPs (color purple in Fig. 6)
and a NdDa component (6 in light brown) (Fig. 8) that demonstrate al-
lelic similarities between East Asians and Neanderthals. However, given
that a component (4) in the overall dataset (Fig. 6) comprises 75.1% of
East Asians (North East Asians and South East Asians), common ances-
tral polymorphisms seem a more likely explanation for the East Asian

Fig. 10. STRUCTURE analysis of 827 contemporary humans, Neanderthals and the Denisova using 4746 SNPs that possess the derived state (different from the chimpanzee) in both the
Neanderthals and the Denisova (NdDd). The number of putative ancestral populations (k) is 10.
 R.K. Lowery et al. / Gene 530 (2013) 83–94 91

Table 2
D-statistic comparison between populations for each SNP subset.

NaDa NdDa NaDd NdDd Population 1 (p1) Population2 (p2)

0 0 0 0 Sub-Saharan Africa Sub-Saharan Africa

−0.14168133 −0.04560491 −0.03056406 −0.00615611 Sub-Saharan Africa Melanesian
−0.14379113 −0.05079328 −0.01322341 0.008469173 Sub-Saharan Africa N.E. Asia
−0.14909445 −0.04975 −0.00915178 0.006670853 Sub-Saharan Africa Amerindian
−0.11342701 −0.04018232 −0.01238406 0.000718881 Sub-Saharan Africa N. African
−0.14784005 −0.05608716 −0.01781661 0.002597354 Sub-Saharan Africa S.W. Asia
−0.14379684 −0.05096687 −0.01709785 0.007268474 Sub-Saharan Africa S. Central Asia
−0.14907196 −0.05477744 −0.01828402 0.004353426 Sub-Saharan Africa S.E. Asia
−0.14973485 −0.05638536 −0.01661412 0.003327806 Sub-Saharan Africa Caucasus
−0.13643782 −0.04912913 −0.01304472 0.004222578 Sub-Saharan Africa Near East
−0.15177493 −0.05772489 −0.01638933 0.003241978 Sub-Saharan Africa European
0.141681329 0.04560491 0.030564065 0.006156111 Melanesian Sub-Saharan Africa
0 0 0 0 Melanesian Melanesian
−0.00215368 −0.00520042 0.017347665 0.014624522 Melanesian N.E. Asia
−0.00757309 −0.00415451 0.021418279 0.012826437 Melanesian Amerindian
0.028715793 0.00543255 0.018186891 0.006874962 Melanesian N. African
−0.00629049 −0.01050913 0.012754396 0.008753326 Melanesian S.W. Asia
−0.00215951 −0.00537445 0.013473259 0.013423985 Melanesian S. Central Asia
−0.00755009 −0.0091955 0.012286908 0.010509256 Melanesian S.E. Asia
−0.00822808 −0.01080824 0.013957033 0.009483723 Melanesian Caucasus
0.005346871 −0.00353213 0.017526334 0.01037842 Melanesian Near East
−0.01031542 −0.01215197 0.014181837 0.009397901 Melanesian European
0.143791128 0.05079328 0.013223411 −0.00846917 N.E. Asia Sub-Saharan Africa
0.002153675 0.005200416 −0.01734767 −0.01462452 N.E. Asia Melanesian
0 0 0 0 N.E. Asia N.E. Asia
−0.00541951 0.001045924 0.004072127 −0.00179842 N.E. Asia Amerindian
0.030867559 0.010632665 0.000839491 −0.00775034 N.E. Asia N. African
−0.00413687 −0.00530901 −0.00459429 −0.00587195 N.E. Asia S.W. Asia
−5.8313E-06 −0.00017404 −0.00387531 −0.00120077 N.E. Asia S. Central Asia
−0.0053965 −0.00399528 −0.00506184 −0.0041159 N.E. Asia S.E. Asia
−0.00607451 −0.00560814 −0.00339145 −0.00514151 N.E. Asia Caucasus
0.00750046 0.001668313 0.000178723 −0.00424675 N.E. Asia Near East
−0.00816193 −0.00695199 −0.00316661 −0.00522734 N.E. Asia European
0.149094447 0.049749998 0.009151777 −0.00667085 Amerindian Sub-Saharan Africa
0.007573092 0.004154514 −0.02141828 −0.01282644 Amerindian Melanesian
0.005419505 −0.00104592 −0.00407213 0.001798422 Amerindian N.E. Asia
0 0 0 0 Amerindian Amerindian
0.036280995 0.009586847 −0.00323265 −0.005952 Amerindian N. African
0.001282665 −0.0063549 −0.00866625 −0.00407357 Amerindian S.W. Asia
0.005413674 −0.00121996 −0.00794731 0.00059765 Amerindian S. Central Asia
2.30031E-05 −0.00504118 −0.00913377 −0.00231749 Amerindian S.E. Asia
−0.00065503 −0.00665403 −0.00746348 −0.00334312 Amerindian Caucasus
0.01291944 0.00062239 −0.00389341 −0.00244834 Amerindian Near East
−0.00274254 −0.00799786 −0.00723864 −0.00342895 Amerindian European
0.113427012 0.040182316 0.012384057 −0.00071888 N. African Sub-Saharan Africa
−0.02871579 −0.00543255 −0.01818689 −0.00687496 N. African Melanesian
−0.03086756 −0.01063266 −0.00083949 0.007750339 N. African N.E. Asia
−0.036281 −0.00958685 0.003232647 0.005952 N. African Amerindian
0 0 0 0 N. African N. African
−0.03499996 −0.01594077 −0.00543376 0.001878477 N. African S.W. Asia
−0.03087339 −0.01080668 −0.00471479 0.006549627 N. African S. Central Asia
−0.03625802 −0.01462732 −0.0059013 0.003634557 N. African S.E. Asia
−0.03693515 −0.01623984 −0.00423093 0.002608932 N. African Caucasus
−0.02337251 −0.00896451 −0.00066077 0.003503708 N. African Near East
−0.03901965 −0.01758336 −0.00400609 0.002523102 N. African European
0.147840055 0.056087162 0.017816614 −0.00259735 S.W. Asia Sub-Saharan Africa
0.006290488 0.010509132 −0.0127544 −0.00875333 S.W. Asia Melanesian
0.004136869 0.005309007 0.004594285 0.005871948 S.W. Asia N.E. Asia
−0.00128267 0.006354896 0.00866625 0.004073569 S.W. Asia Amerindian
0.034999959 0.015940772 0.005433755 −0.00187848 S.W. Asia N. African
0 0 0 0 S.W. Asia S.W. Asia
0.004131037 0.005134972 0.000718986 0.004671208 S.W. Asia S. Central Asia
−0.00125966 0.001313758 −0.00046756 0.001756092 S.W. Asia S.E. Asia
−0.00193769 −0.00029915 0.001202851 0.000730458 S.W. Asia Caucasus
0.011636968 0.006977258 0.004773004 0.001625242 S.W. Asia Near East
−0.00402519 −0.00164305 0.001427699 0.000644629 S.W. Asia European
0.143796839 0.050966869 0.017097847 −0.00726847 S. Central Asia Sub-Saharan Africa
0.002159507 0.005374451 −0.01347326 −0.01342398 S. Central Asia Melanesian
5.83134E-06 0.00017404 0.003875312 0.001200773 S. Central Asia N.E. Asia
−0.00541367 0.001219964 0.007947314 −0.00059765 S. Central Asia Amerindian
0.030873385 0.010806685 0.004714788 −0.00654963 S. Central Asia N. African
−0.00413104 −0.00513497 −0.00071899 −0.00467121 S. Central Asia S.W. Asia
0 0 0 0 S. Central Asia S. Central Asia
−0.00539067 −0.00382124 −0.00118655 −0.00291514 S. Central Asia S.E. Asia

(continued on next page)

92 R.K. Lowery et al. / Gene 530 (2013) 83–94

Table 2 (continued)
NaDa NdDa NaDd NdDd Population 1 (p1) Population2 (p2)

−0.00606868 −0.00543411 0.000483865 −0.00394076 S. Central Asia Caucasus

0.007506291 0.001842353 0.004054033 −0.00304599 S. Central Asia Near East
−0.00815609 −0.00677796 0.000708714 −0.00402659 S. Central Asia European
0.149071956 0.054777441 0.018284023 −0.00435343 S.E. Asia Sub-Saharan Africa
0.00755009 0.009195502 −0.01228691 −0.01050926 S.E. Asia Melanesian
0.005396503 0.003995277 0.005061836 0.004115898 S.E. Asia N.E. Asia
−2.3003E-05 0.005041181 0.009133774 0.002317494 S.E. Asia Amerindian
0.036258022 0.014627321 0.005901302 −0.00363456 S.E. Asia N. African
0.001259662 −0.00131376 0.000467561 −0.00175609 S.E. Asia S.W. Asia
0.005390672 0.00382124 0.001186547 0.00291514 S.E. Asia S. Central Asia
0 0 0 0 S.E. Asia S.E. Asia
−0.00067803 −0.0016129 0.001670411 −0.00102563 S.E. Asia Caucasus
0.012896441 0.005663552 0.005240554 −0.00013085 S.E. Asia Near East
−0.00276554 −0.0029568 0.001895259 −0.00111146 S.E. Asia European
0.149734854 0.056385362 0.016614119 −0.00332781 Caucasus Sub-Saharan Africa
0.008228082 0.010808245 −0.01395703 −0.00948372 Caucasus Melanesian
0.006074514 0.005608144 0.003391454 0.005141511 Caucasus N.E. Asia
0.00065503 0.00665403 0.007463477 0.003343121 Caucasus Amerindian
0.036935148 0.016239841 0.004230933 −0.00260893 Caucasus N. African
0.001937694 0.000299146 −0.00120285 −0.00073046 Caucasus S.W. Asia
0.006068683 0.00543411 −0.00048387 0.003940763 Caucasus S. Central Asia
0.000678033 0.001612903 −0.00167041 0.001025635 Caucasus S.E. Asia
0 0 0 0 Caucasus Caucasus
0.013574356 0.007276389 0.003570174 0.000894784 Caucasus Near East
−0.00208752 −0.0013439 0.000224849 −8.583E-05 Caucasus European
0.136437816 0.04912913 0.013044719 −0.00422258 Near East Sub-Saharan Africa
−0.00534687 0.003532133 −0.01752633 −0.01037842 Near East Melanesian
−0.00750046 −0.00166831 −0.00017872 0.004246746 Near East N.E. Asia
−0.01291944 −0.00062239 0.003893407 0.002448344 Near East Amerindian
0.02337251 0.008964511 0.000660768 −0.00350371 Near East N. African
−0.01163697 −0.00697726 −0.004773 −0.00162524 Near East S.W. Asia
−0.00750629 −0.00184235 −0.00405403 0.003045989 Near East S. Central Asia
−0.01289644 −0.00566355 −0.00524055 0.000130851 Near East S.E. Asia
−0.01357436 −0.00727639 −0.00357017 −0.00089478 Near East Caucasus
0 0 0 0 Near East Near East
−0.01566143 −0.00862021 −0.00334533 −0.00098061 Near East European
0.151774929 0.057724889 0.016389332 −0.00324198 European Sub-Saharan Africa
0.010315419 0.012151969 −0.01418184 −0.0093979 European Melanesian
0.008161926 0.006951993 0.003166607 0.005227339 European N.E. Asia
0.002742542 0.007997859 0.007238641 0.003428949 European Amerindian
0.039019654 0.017583358 0.004006088 −0.0025231 European N. African
0.004025193 0.001643046 −0.0014277 −0.00064463 European S.W. Asia
0.008156095 0.006777961 −0.00070871 0.004026591 European S. Central Asia
0.002765545 0.002956798 −0.00189526 0.001111465 European S.E. Asia
0.002087515 0.001343901 −0.00022485 8.58298E-05 European Caucasus
0.015661427 0.008620206 0.003345328 0.000980614 European Near East
0 0 0 0 European European

affinity to Neanderthals, especially if we consider that Neanderthals'
geographical distribution range is not known to include East Asia.
4. Conclusions
Using PCA and Structure analyses, we observed genetic sharing be-
tween Europeans and Neanderthals that may be due to admixture
events or ancestral common polymorphisms. Further, if this sharing is
due to common ancestral polymorphisms rather than introgression, ei-
ther genetic drift demic diffusion or ancestral population substructure
may explain the greater affinity of Neanderthals to Europeans as
compared to other Eurasians to the east. Although the observed clin-
al degree of similarities seen in the Structure analysis could result
from introgression followed by demic diffusion, the data is easier
explained by common ancestral polymorphisms shared between the
ancestors of Modern Europeans and Neanderthals, followed by demic
diffusion eastward. In particular, if Neanderthal admixture with
modern humans is responsible for the 65.4% (component 7, Fig. 6)
of the European nuclear genome shared with Neanderthals, it con-
trasts the complete lack of Neanderthal contribution to the European
mitochondria or y-chromosomes (Herrera et al., 2009), yet, it is pos-
sible that introgression may have contributed in some degree to the
observed genetic affinity. Although uniparental dropoff could have
erased evidence of Neanderthal contribution to the modern human
gene pool, it is unlikely that no abundant Neanderthal genetic signals
would be detected in nuclear DNA of contemporary human populations.
However, if gene flow is the source of the European/Neanderthal genetic
commonalities, it would appear more likely that humans are the source
of the flow, as the shared component is more characteristic of Europeans
(65.4%) than Neanderthals (3.6%). Recent dating of putative Neanderthal
haplotypes in Europeans indicates an entrance into the European gene
pool of between 37,000 and 86,000 years before present (Sankararaman
et al., 2012). Yet, these data do not assess whether these dates reflect
Neanderthal–modern human introgression or represent signals from
an ancient modern human population with a closer kinship with
Neanderthals via ancestral African subpopulation structure.
An out of Africa hypothesis posits that difficult climatic conditions
restricted an early migration of humans, approximately 70 kya, along
the southern coast of Asia. Based on this hypothesis, the Melanesians
are thought to descend from a population of modern humans that
reached Melanesia earlier than the peopling of most Eurasia. As a result,
it is possible that Melanesians have retained more genetic material
from the common ancestral gene pool than Eurasians. Therefore, the
D-statistic-based results indicating that Melanesians possess a greater
number of NaDd and NdDd alleles is more congruent with a model in
which common ancestral polymorphisms followed by genetic drift
 R.K. Lowery et al. / Gene 530 (2013) 83–94
Fig. 11. D-statistics (Green et al., 2010) are graphed individually for each of the 11 contemporary human metapopulations using each of the 4 SNP subsets. X axis represents specific
metapopulations. Y axis indicate D-statistic values. For key to abbreviations indicating populations included in each metapopulation group see Table 1.
during the out of African diaspora led to Denisova–Melanesian similar-
ities rather than recent admixture.
In summary, the currently available information from previous pub-
lished work as well as the results from the present study indicate that
the observed genetic affinities between archaic and contemporary
human populations are mostly explained by ancestral common poly-
morphims followed by genetic drift, and not admixture, as the most
parsimonious model. This, of course, does not rule out the possibility
of introgression events contributing to some degree to the archaic–
contemporary human commonalities. Continued sequencing efforts,
combined with the development of novel approaches to phylogenetic
and genetic structure analyses, may hold the key to further unraveling
some of these unanswered questions.
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.gene.2013.06.005.
Conflict of interest
None.
References
Behar, D.M., et al., 2010. The genome-wide structure of the Jewish people. Nature 466,
238–242.
Cann, R.L., 2001. Genetic clues to dispersal in human populations, retracing the past from
the present. Science 291, 1742–1748.
Cavalli-Sforza, L.L., Menozzi, P., Piazza, A., 1994. The History and Geography of Human
Genes. Princeton University Press.
Durand, E.Y., Patterson, N., Reich, D., Slatkin, M., 2011. Testing for ancient admixture
between closely related populations. Mol. Biol. Evol. 28 (8), 2239–2252.
EIGENSOFT software package 2006. [Internet]. Boston, Reich Laboratory, Harvard Medical
School, Department of Genetics. Available, http://genepath.med.harvard.edu/~reich/
Software.htm. Accessed 1 March 2012.
93
Eriksson, A., Manica, A., 2012. Effect of ancient population structure on the degree of poly-
morphism shared between modern human populations and ancient hominins. Proc.
Natl. Acad. Sci. 109, 13956–13960.
Green, R.E., et al., 2010. A draft sequence of the Neanderthal genome. Science 328,
710–722.
Herrera, K.J., Somarelli, J.A., Lowery, R.K., Herrera, R.J., 2009. To what extent did Neanderthals
and modern humans interact? Biol. Rev. Camb. Philos. Soc. 84 (2), 245–257.
Jakobsson, M., et al., 2008. Genotype, haplotype and copy-number variation in worldwide
human populations. Nature 451, 998–1003.
Kennedy, K.A.R., Deraniyagala, S.U., Roertgen, W.J., Chiment, J., Disotell, T., 1987. Upper
Pleistocene fossil hominids from Sri Lanka. Am. J. Phys. Anthropol. 72 (4).
Kong, Q.P., Bandelt, H.J., Sun, C., Yao, Y.G., Salas, A., Achilli, A., et al., 2006. Updating the
East Asian mtDNA phylogeny: a prerequisite for the identification of pathogenic mu-
tations. Hum. Mol. Genet. 15 (13), 2076–2086.
Lahr, M.M., Foley, R., 1994. Multiple dispersals and modern human origins. Evol.
Anthropol. Issues News Rev. 3 (2).
Langergraber, K.E., Prüfer, K., Rowney, C., Boesch, C., Crockford, C., Fawcett, K., et al., 2012.
Generation times in wild chimpanzees and gorillas suggest earlier divergence times
in great ape and human evolution. Proc. Natl. Acad. Sci. 109 (39), 15716–15721.
Li, J.Z., et al., 2008. Worldwide human relationships inferred from genome-wide patterns
of variation science. 319, 1100–1104.
Maca-Meyer, N., González, A.M., Larruga, J.M., Flores, C., Cabrera, V.M., 2001. Major
genomic mitochondrial lineages delineate early human expansions. BMC Genet. 2,
1471–2156.
Macaulay, V., et al., 2005. Single, rapid coastal settlement of Asia revealed by analysis of
complete mitochondrial genomes. Science 308, 1034–1036.
Meyer, M., et al., 2012. A high-coverage genome sequence from an archaic Denisovan
individual. Science 338, 222–226.
Palanichamy, M., et al., 2004. Phylogeny of mitochondrial DNA macrohaplogroup N in
India, based on complete sequencing: implications for the peopling of South Asia.
Am. J. Hum. Genet. 75, 966–978.
Patterson, N., Price, A.L., Reich, D., 2006. Population structure and eigenanalysis. PLoS
Genet. 2, e190.
Price, A.L., Patterson, N.J., Plenge, R.M., Weinblatt, M.E., Shadick, N.A., Reich, D., 2006. Prin-
cipal components analysis corrects for stratification in genome-wide association
studies. Nat. Genet. 38 (8), 904–909.
Quintana-Murci, L., Semino, O., Bandelt, H.J., Passarino, G., McElreavey, K., Santachiara-
Benerecetti, A.S., 1999. Genetic evidence of an early exit of Homo sapiens sapiens
from Africa through eastern Africa. Nat. Genet. 23, 437–441.
Reich, D., et al., 2010. Genetic history of an archaic hominin group from Denisova Cave in
Siberia. Nature 468, 1053–1060.
94 R.K. Lowery et al. / Gene 530 (2013) 83–94

Sankararaman, S., Patterson, N., Li, H., Pääbo, S., Reich, D., 2012. The date of inter-
breeding between Neandertals and modern humans. PLoS Genet. 8 (10),
e1002947.
Skoglund, P., Jakobsson, M., 2011. Archaic human ancestry in East Asia. Proc. Natl. Acad.
Sci. U. S. A. 108, 18301–18306.
Slatkin, M., Pollack, J., 2008. Subdivision in an ancestral species creates asymmetry in gene
trees. Mol. Biol. Evol. 25 (10), 2241–2246.
Stringer, C., 2000. Palaeoanthropology coasting out of Africa. Nature 405, 24–27.
Sun, C., et al., 2006. The dazzling array of basal branches in the mtDNA macrohaplogroup
M from India as inferred from complete genomes. Mol. Biol. Evol. 23, 683–690.
Tanaka, M., et al., 2004. Mitochondrial genome variation in eastern Asia and the peopling
of Japan. Genome Res. 14, 1832–1850.
Wall, J.D., et al., 2013. Higher levels of Neanderthal ancestry in East Asians than in
Europeans. Genetics 112, 148213.
Yang, M.A., Malaspinas, A.S., Durand, E.Y., Slatkin, M., 2012. Ancient structure in Africa unlikely to
explain Neanderthal and Non-African genetic similarity. Mol. Biol. Evol. 29 (10), 2987–2995.