AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 000:000–000 (2006)

Mitochondrial DNA Analysis of Ancient

Peruvian Highlanders
Ken-ichi Shinoda,1* Noboru Adachi,2 Sonia Guillen,3 and Izumi Shimada4
1
Department of Anthropology, National Science Museum, Tokyo 169-0073, Japan
2
Department of Anatomy and Anthropology, School of Medicine, Tohoku University, Sendai 980-8575, Japan
3
Centro Mallqui Institute, Lima 11, Peru
4
Department of Anthropology, Southern Illinois University, Carbondale, Illinois 62901

KEY WORDS Inca; ancient DNA; Urubamba; Peru

ABSTRACT Ancient DNA recovered from 57 individu- and Bolivian highlands, and contrast with those of pre-
als excavated by Hiram Bingham at the rural communities Hispanic individuals of the north coast of Peru that we
of Paucarcancha, Patallacta, and Huata near the famed defined previously. Our study suggests a strong genetic
Inca royal estate and ritual site of Machu Picchu was affinity between sampled late pre-Hispanic individuals and
analyzed by polymerase chain reaction, and the results modern Andean highlanders. A previous analysis of the
were compared with ancient and modern DNA from vari- Machu Picchu osteological collection suggests that the resi-
ous Central Andean areas to test their hypothesized indig- dents there were a mixed group of natives from various
enous highland origins. The control and coding regions of coastal and highland regions relocated by the Inca state
the mitochondrial DNA (mtDNA) of 35 individuals in this for varied purposes. Overall, our study indicates that the
group were sequenced, and the haplogroups of each indi- sampled individuals from Paucarcancha and Patallacta
vidual were determined. The frequency data for the hap- were indigenous highlanders who provided supportive
logroups of these samples show clear proximity to those of roles for nearby Machu Picchu. Am J Phys Anthropol 000:
modern Quechua and Aymara populations in the Peruvian 000–000, 2006. V 2006 Wiley-Liss, Inc.
C

The discovery of the magnificent site of Machu Picchu This study suggests that Patallacta was an important
in 1911, at an elevation of ca. 2,740 m above sea level Inca local administrative center, surrounded by satellite
atop a ridge connecting two granite peaks in the Eastern settlements with varied functions that complemented
Cordillera of the Peruvian Andes by Bingham (1912, the center. Patallacta had 115 kanchas or walled com-
1913), generated great professional and public interest in pounds, each composed of four buildings surrounding a
the site and in the Inca Empire in general. It also brought central courtyard, that were organized into ‘‘four distinct
attention to the prehistory of this heretofore little-known residential sectors offering accommodation for social
rugged area approximately 100 km (ca. a 3-day walk) to groups of varied status and structure’’ (Kendall, 1988).
the northwest of the Inca capital of Cuzco. This settlement complex included the relatively small
Understandably, the discovery of Machu Picchu raised but well-preserved ‘‘fort’’ of Paucarcancha, the place from
broader questions, such as whether the settlement was which the majority of human remains that were ana-
isolated, intrusive, and/or specialized; and if not, the loca- lyzed in the present study were derived. The site is
tions of the contemporaneous settlements in which the located approximately 5 km upstream (south) of Patal-
sustaining population lived; and whether the residents lacta, atop terraces (3,000 m above sea level) overlooking
were local people. The explorations by Bingham in 1911 the confluence of the Cusichaca and the Quescamayo
and 1912 focused on Machu Picchu and only identified a Rivers; it presumably guarded the road to and from the
handful of relatively accessible sites on the Inca road to Urubamba Valley (Kendall, 1985). The curving perimeter
it. Bingham (1916, 1930) returned to the vicinity of Machu wall encloses approximately 18 rectangular structures.
Picchu in 1914 (spanning 1914–1915) and attempted to Based on its location relative to various major state
answer these questions. Though the survey that aimed at
locating additional Inca sites was conducted along Inca
roads (Fig. 1), his fieldwork focused mainly on the excava-
Grant sponsor: Ministry of Education, Science, Sports and Culture,
tion of the site at Patallacta (also known as Lacta Pata, ca. Japan; Grant number: 13575017.
2,500 m above sea level) that he had explored and mapped
in 1911, prior to finding Machu Picchu. This site is impres- *Correspondence to: Ken-ichi Shinoda, Department of Anthropology,
sive with respect to its size, orderly planned layout, and National Science Museum, 3-23-1 Hyakunincho, Shinjyuku-ku, Tokyo
placement on an artificially leveled platform. The platform 169-0073, Japan. E-mail: shinoda@kahaku.go.jp
was carved out of a semicircular promontory overlooking
the confluence of the Cusichaca and the Urubamba Rivers, Received 18 February 2005; accepted 21 October 2005.
a 17-km linear distance to the southeast of Machu Picchu.
A more recent and comprehensive study of this site and DOI 10.1002/ajpa.20408
nearby Inca sites was conducted by the Cusichaca Published online in Wiley InterScience
Archaeological Project led by Kendall (1974, 1985, 1988). (www.interscience.wiley.com).

C 2006
V WILEY-LISS, INC.
2 K. SHINODA ET AL.

Fig. 1. Map of Machu Picchu

and other archaeological sites.
Circles indicate sites discovered
by Hiram Bingham. Solid circles
indicate sites analyzed in this
study.

installations and roads and its architectural and ceramic Inca-period burials at surveyed settlements near Machu
styles, the history of Paucarcancha dates back to the reign Picchu may include both local and genetically distinct non-
of the Inca king Topa Inca (son of the king Pachacuti local individuals.
Inca Yupanqui), approximately in the late 15th century The present study aims at shedding light on the bio-
(Kendall, 1985). Based on architecture, ceramics, and other logical identities and geographical origins of Inca-period
artifacts found in association, the burials that Bingham residents of the Machu Picchu-Cusichaca section of the
excavated at Paucarcancha and Patallacta can be assigned Urubamba Valley. Eaton (1916) described the materials
to the period of the Inca control of the Urubamba Valley, recovered at Machu Picchu by the projects led by Hiram
from ca. mid-15th to early 16th centuries (Bingham, 1913; Bingham of Yale University. He included a thorough re-
Kendall, 1985; MacCurdy, 1923). port on the bones and artifacts, as well as information
Over the past 20 years, in addition to the aforemen- on the condition of the caves where the material had
tioned work led by Kendall, there has been much effort been located. Eaton (1916) concluded that females greatly
to elucidate Inca and pre-Inca occupations along the ‘‘Sa- outnumbered males at Machu Picchu, by a ratio of approx-
cred Valley of the Inca’’ and further downstream along imately 4:1. This is the scientific basis for Bingham’s idea
the Urubamba River past Machu Picchu (e.g., Farring- that the site included an Aclla Wasi, a House of the
ton, 1995; Niles, 1984, 1988, 1999; Protzen, 1993). Bing- Chosen Women. More recent research questioned this
ham (1930) believed that Machu Picchu was built and interpretation. Reanalysis of the data of Eaton (1916)
occupied by various Inca kings and served as a citadel revealed that the sex ratio was 1.54:1 (Verano, 2003).
and as the final Inca capital of Vilcapampa. Eaton (1916) also proposed that the presence of brachy-
On the other hand, following a report by Rowe (1990), cephalic and dolicocephalic skulls was indicative of high-
a majority of scholars now believe that it was a royal land and coastal groups. The recent examination of vari-
country estate or a palace of the Inca king Pachacuti Inca ous types of cranial deformities, and a related multivariate
Yupanqui. In fact, Machu Picchu was one of at least craniometric analysis of the Machu Picchu osteological col-
18 palaces built by three generations of Inca kings (Niles, lection by Verano (2003), suggest that the residents were a
1988, 1999). mixed group of natives from various coastal and highland
These estates were associated with large-scale, state- regions. These findings support the earlier conclusion by
directed land-reclamation and land-management projects Eaton (1916) and raise an intriguing question regarding
that created thousands of irrigated agricultural terraces, whether the residents of nearby communities were also of
making the Urubamba Valley the breadbasket for the mixed origins.
Inca capital of Cuzco (e.g., Kendall, 1988). In addition to MacCurdy (1918, 1923), the physical anthropologist
its relatively low elevation, as Protzen (1993) noted, the who participated in the aforementioned Yale University
Sacred Valley with its roughly east-west orientation cre- expedition of 1914–1915, studied the material recovered
ated a thermal belt or microenvironment that favored in- by Bingham from places near Machu Picchu, i.e., Paucarcan-
tensive agriculture. Large-scale maize cultivation at that cha, Patallacta, Torony, Sillque, Huispand, Huata, Huaro-
site appears to have been manned, in part, by ‘‘experi- condo, and Yanamanchi (Fig. 1). MacCurdy (1918, 1923)
enced maize growers from other parts’’ of the Inca Empire also made thorough descriptions of the material studied,
(Von Hagen and Morris, 1998; Niles, 1984). A majority of including statistical information. A series of measure-
the 68 settlements in the Inca-period settlements of di- ments was included, along with photographs of the sites
verse sizes that Kendall (1985, 1988) recorded (including and bones. His summary of the study of the material from
Patallacta and Torontoy) along three major roads between the eight sites indicated that according to the funerary
the Santa Teresa (downstream from Machu Picchu) and pattern and characters, the individuals all belonged to
Anta Rivers (just upstream from Ollantaytambo) probably the same group. They had been buried in caves and shel-
served, in part, as residential settlements for both local ters. He noted the high frequency of the highland or
and relocated farmers. These observations suggest that Aymara type of deformation and the almost complete ab-
 ANCIENT DNA ANALYSIS OF PERUVIAN HIGHLANDERS 3
sence of the brachycephalic, coastal type. This was in Authentication methods
great contrast to the Machu Picchu sample.
As the Inca state practiced forced relocation of popula- No modern DNA-based studies had been performed
tions for political-military, economic, and other purposes previously in the area subsumed by our ancient DNA
(Rowe, 1982), there is a question as to the provenience study. Standard contamination precautions, such as sep-
and biological identities of residents at the sites from aration of pre- and post-polymerase chain reaction (PCR)
which the samples were derived. While ethnohistorical experimental areas, use of disposable laboratory ware
documents and architectural quality and style, as well and filter-plugged pipette tips, treatment with DNA
as site location, all point to Machu Picchu as a planned contamination removal solution (DNA-OFFTM, TaKaRa,
royal estate and ceremonial center, the connected settle- Otsu, Japan), ultraviolet irradiation of equipment and
ments of Paucarcancha, Patallacta, and Huata lack ar- benches, negative extraction controls, and negative PCR
chitectural or other features diagnostic of Inca royal or controls, were employed in the present study. DNA-based
ceremonial sites; rather, they show similarities to local experiments were performed by two authors (N.A. and
pre-Inca antecedents. Based on these observations and K.-I.S.) independently, in different laboratories for cross-
their proximity to cultivable fields and water sources, we validation.
hypothesize that the burials excavated at Paucarcancha
and Patallacta are those of highland agricultural popula-
tions derived from immediate or nearby highland areas. Extraction and purification of DNA
Accordingly, we expect that these burials as a group
would be characterized by relatively homogeneous DNA Tooth samples were dipped in a 13% bleach solution
composition, similar to modern-day highland Indians for 15 min, rinsed several times with DNase-/RNase-free
(Quechua speakers). On the other hand, the presence of distilled water (Invitrogen, Carlsbad, CA), and allowed
individuals relocated from distant lands within the Inca to air-dry. Moreover, using a dental drill, the outer sur-
Empire would yield an internally heterogenous genetic face of samples was removed to a depth of 1 mm. Next,
composition. We would expect a contrasting situation for samples were again rinsed with DNase-/RNase-free dis-
the Machu Picchu. tilled water and allowed to air-dry. After samples were
In the current study, we present data from mtDNA con- completely dry, they were irradiated in an ultraviolet
trol and coding region of ancient Peruvian highlanders in cross-linker for 30 min and turned over several times.
the Urubamba Valley, to address questions about the Next, samples were pulverized in a mill (Multi-beads
genetic relationships among these ancient groups and Shocker MB400U, Yasui Kikai, Osaka, Japan). The pow-
South American populations. Our purpose is to test the dered samples (0.3–0.5 g) were decalcified with 0.5 M
hypothesis that ancient highlanders were more closely
EDTA (pH 8.0) (Invitrogen) at 48C for 5–7 days, and
related to modern highlanders than to other modern pop-
then rinsed three times with DNase-/RNase-free distilled
ulations in the region.
water.
Two different methods were employed for the extrac-
tion of DNA from the powdered samples. In the first
MATERIALS AND METHODS method, decalcified samples were lysed in 3 ml of Buffer
Sampling ASL (Qiagen, Hilden, Germany) with 150 ll of 20 mg/ml
Proteinase K (Invitrogen) at 558C overnight. DNA was
To analyze the mitochondrial DNA of these residents, extracted from the lysate using a QIAamp1 DNA Stool
we sampled 57 teeth that had been collected by G. Mac- Mini Kit (Qiagen), in accordance with the technical man-
Curdy. We sampled 26 individuals from Paucarcancha,
ual. In the second method, DNA was extracted in two
24 from Patallacta, and seven from Huata. Although the
steps, using a GENECLEAN kit for ancient DNA (Bio
materials from this expedition had been initially ex-
ported to the US, they were returned to Peru (Burger 101). The pulverized tooth (powder) was placed in a 15-
and Salazar, 2003) ‘‘in accordance with the original agree- ml conical tube, and 5 ml of an overnight soaking solu-
ment with the Peruvian government,’’ and are housed tion were added for Proteinase K digestion. Samples
in the National Museum of Anthropology, Archaeology were rotated and incubated at 378C for 12–15 hr. The su-
and History in Lima (Museo Nacional de Antoropologia pernatant was used for DNA extraction with the kit.
Arqueologia e Historia del Peru, MNAAH). Approximately 100 ll of extracted DNA solution were
Because the collection was returned to Peru, these obtained. The eluted DNA was amplified by PCR without
skeletal remains were preserved commingled with mate- further processing. DNA extraction was done only once,
rials from other sites. The remains were sorted out in ac- and if the following PCR amplification was not success-
cordance with the provenience information recorded in ful, no further extraction was carried out.
the Museum catalog. However, we were able to find sam-
ples recovered from only three sites, i.e., Paucarcancha,
Patallacta, and Huata. There is a possibility that the Amplification and sequencing of HVR 1,
remaining samples are lying unrecorded in the MNAAH HVR 2, and coding region 10382–10465
without any record. Only specimens that were clearly la-
beled and that coincided with Macurdy’s inventory num- A segment of hypervariable region (HVR) 1 (nucleotide
bers were used in our study. positions 16209–16402, relative to the revised Cambridge
Teeth were collected only from the mandible, to avoid reference sequence (CRS); Andrews et al., 1999), HVR 2
the possibility of sampling from the same individual. A (128–267), and a segment of the coding region (10382–
well-preserved tooth was extracted from each individual. 10465) that covers a part of the NADH dehydrogenase 3
Teeth, thus collected, were exported from Peru to Japan and tRNAArg gene were sequenced for all samples. The
with the permission of the National Institute of Culture mtDNA sequence can be tentatively assigned to respec-
of Peru. tive haplogroups according to specific mutations ob-
4 K. SHINODA ET AL.
TABLE 1. Primers used for sequence analysis
Nucleotide Annealing
Primer Sequence (50 –30 ) positions temperature (8C) Reference
M13-L127 ( 21M13)1 AGCACCCTATGTCGCAGTAT 128–267 46 Adachi et al. (2003)
M13-H268 (M13 reverse)2 TGTTATGATGTCTGTGTGG
M13-L10381 ( 21M13) CTGGCCTATGAGTGACTACA 10382–10465 50 Adachi et al. (2004)
M13-H10466 (M13 reverse) TGTAAATGAGGGGCATTTGG
M13-L16208 ( 21M13) CCCCATGCTTACAAGCAAG 16209–16402 50 Nata et al. (1999)
M13-H16403 (M13 reverse) ATTGATTTCACGGAGGATGG
1
21M13 sequence represents 50 -TGTAAAACGACGGCCAGT-30 .
2
M13 reverse sequence represents 50 -AACAGCTATGACCATG-30 .

Fig. 2. APLP band patterns of diagnostic regions for haplogroups A, B, C, and D. Lane N indicates that no nucleotide change
was observed. Lane M indicates mutation at site. Lane LM indicates 10-bp ladder marker (Invitrogen).

served in the HVR 1 region. Further characterization of Amplified product-length

haplogroup status was by means of other specific muta- polymorphisms analysis
tions in HVR 2 and the coding region. This coding region
sequence includes the detection site of macrohaplogroups Several methods are available for the detection of sin-
M and N. Each of sites 10398 and 10400 is one of the gle-nucleotide polymorphisms (SNPs). Restriction frag-
defining sites for macrohaplogroups M and N (Quintana- ment-length polymorphism (RFLP) analysis is one of the
Marci et al., 1999). The known Native American hap- most widely used techniques for detecting known muta-
logroups A and B are included in N, while C and D tion sites in the coding region of mtDNA. Recently, a
pertain to M. Detection of this coding region provides a simple and rapid amplified product-length polymor-
phisms (APLP) method was developed for the analysis of
firmer basis for haplogroup assignment. If haplogroup
mtDNA polymorphisms (Umetsu et al., 2001). The prin-
assignments derived from HVR 1 and coding region
ciple of this method is based on an attachment of a non-
sequences did not accord with each other, the result was
complementary sequence to the 50 -end of one of two al-
rejected. Primers used to amplify the regions described lele-specific primers. The use of such primers permits
above are listed in Table 1. the amplification of two size-different products distin-
A 1-ll aliquot of the extract was used as the template guishing two alleles (Fig. 2). Using APLP, diagnostic poly-
for PCR. Amplifications were carried out in a total reac- morphic sites can be examined directly. Therefore, ambi-
tion volume of 10 ll containing 0.25 units of Taq DNA guity with respect to the actual mutation site, which is
polymerase (HotStarTaqTM DNA polymerase, Qiagen), one of the problems of the RFLP method caused by
0.2 lM of each primer, 200 lM of dNTPs, and 0.1 lg/ll insufficient enzymic digestion, can be avoided.
of bovine serum albumin (Amersham Pharmacia Biotech, It is generally agreed that most of the mtDNA from
Uppsala, Sweden) in 1 3 PCR buffer (Qiagen). The con- Native Americans could be traced to one of four mater-
ditions for PCR included incubation at 958C for 15 min; nal lineages present in the founders of New World popu-
40 cycles at 948C for 20 sec, 468C–508C for 20 sec, and lations. Further, these lineages can be defined by three
728C for 15 sec; and final extension at 728C for 1 min. restriction-site polymorphisms and a 9-bp deletion (Wal-
Following PCR, the products were purified with a lace et al., 1985). In the present study, three SNPs in
MinEluteTM PCR Purification Kit (Qiagen) and subjected the coding region (np 5178, 8794, and 14318) and a 9-bp
to direct sequencing. Sequence reactions were prepared repeat variation in the noncoding cytochrome oxidase II/
using 21M13 and M13 reverse primers and a DYEnamic tRNALys intergenic region (9 bp) were analyzed. The pri-
ET Terminator Cycle Sequencing Kit (Amersham Phar- mers used for analyses are listed in Table 2. The consti-
macia Biotech). All sequencing reactions were analyzed tution of the PCR reaction mixture was the same as
using a 377 or 310 DNA Sequencer with SeqEd software described above. Thermal conditions were as follows:
(Applied Biosystems, Foster City, CA), and the sequence incubation at 958C for 15 min; 40 cycles at 948C for 10
of each region that did not contain primer regions was sec, 528C–548C for 10 sec, and 728C for 5 sec; and final
determined and compared with the revised CRS. extension at 728C for 1 min. Each region was examined
 ANCIENT DNA ANALYSIS OF PERUVIAN HIGHLANDERS 5
TABLE 2. Primers used for APLP analysis
Product size Annealing
Site Primer Sequence1 (50 –30 ) (genotype) temperature (8C) Reference
5178 5178C gTCGCACCTGAAgCAAGC 96 bp (C type) 52 Umetsu et al. (2001)
5178A tgatcaaCGCACCTGAAACAAGA 101 bp (A type)
5178R attGCAAAAAGCaGGTTAGCG
8794 8794C aCCTCGGACTtCTGCCTC 107 bp (C type) 52 Umetsu et al. (2001)
8794T attggaCCTCGGACTCtTGCCTT 112 bp (T type)
8794R aCAGCGAAAGCCTATAATCAC
14318 14318T CCTTCATAAATTATTCAGCTTCCaACACTAT 110 bp (T type) 54 Present study
14318C aaaaagctaCATAAATTATTCAGCTTCCTACtCTAC 115 bp (C type)
14318R TTAGTGGGGTTAGCGATGGA
9 bp 9bp-F (-21M13)2 ACAGTTTCATGCCCATCGTC 143 bp (1; deletion) 50 Present study
9bp-R (M13 reverse)3 CTAAGTTAGCTTTACAGTGGG 152 bp (2; normal)
1
Noncomplementary nucleotides are written in lower-case letters.
2
-21M13 sequence represents 50 -TGTAAAACGACGGCCAGT-30 .
3
M13 reverse sequence represents 50 -AACAGCTATGACCATG-30 .

independently, using the monoplex PCR method to maxi- Population differentiation tests (Raymond and Rous-
mize the robustness of PCR. set, 1995) were computed by means of the Arlequin
A 1-ll aliquot of the PCR product was separated by 2.000 computer program (Schneider et al., 2000).
electrophoresis in an 8-cm native polyacrylamide gel (10%
T, 5% C) containing 1 3 TBE buffer (pH 8.0) with running
buffer (0.5 3 TBE, pH 8.0). DNA bands were detected RESULTS
by ultraviolet irradiation after staining with ethidium
bromide (Fig. 2). Throughout the study, both negative extraction con-
trols and negative PCR controls showed negative results
Data analysis consistently. Of the 57 specimens examined, 35 speci-
mens were successfully analyzed (Table 3). The first
With improved knowledge of the global mtDNA tree in method that used the stool kit recovered DNA from 31
recent years, an understanding of the structure of samples. On the other hand, the second method that
mtDNA data and assigning the mtDNA type to a place used the ancient DNA kit recovered DNA from 35 sam-
in the global mtDNA tree have been simplified. Control- ples. Thus, the ancient DNA kit was observed to be
region motifs were identified for a majority of the major the superior method for recovering DNA. However, sus-
haplogroups and their subhaplogroups (Alves-Silva et al., pected false-positive results stemming from contamina-
2000; Bandelt et al., 2001; Kivisild et al., 2002; Kong tion with contemporary DNA and other questionable data
et al., 2003; Macaulay et al., 1999; Maruyama et al., (e.g., Kolman and Tuross, 2000) were obtained while using
2003; Quintana-Murci et al., 1999; Yao et al., 2002, 2003). this method. Following this criterion, two of the sequences
Therefore, we assigned each mtDNA to haplogroups were omitted.
according to the HVR 1, HVR 2, and coding-region data, A majority of the mtDNA of these specimens fully rep-
using the data and classification tree described above, resented haplogroup motifs; therefore, we could safely
such that each sample was allocated to the smallest assign them to the relevant haplogroups. This implied
named haplogroup to which it belonged. If the haplogroup that the authenticity of data obtained from these speci-
had further characterized subhaplogroups, an asterisk mens was supported by established, reliable, modern
was attached to the name of the haplogroup to indicate human mtDNA data. However, in the DNA analysis of
that the haplogroup status could not be identified further ancient samples, the possibility that the original sequen-
(Table 3). Since several segments of the same mtDNA ces may have changed due to postmortem damage to the
were analyzed independently, meticulous care was taken DNA (Thomas et al., 2003) must also be considered.
to avoid artificial recombination caused by potential sam- Thus, it is not advisable to assume the accuracy of all
ple crossover. After assigning the mtDNAs to relevant hap- the base sequences determined in the investigation
logroups, we classified them further into maternal lines, under discussion. It should be clearly understood that
based on the nucleotide changes observed in the control such a limitation would inevitably occur in the analyses
and coding regions. of the scarce DNA that remains in the ancient samples.
To elucidate biological relationships between hypothe- In fact, the sequence of some nucleotide positions remained
sized ancient highlanders from the communities of Pau- uncertain in some specimens (Table 3).
carcancha, Patallacta, and Huata on the one hand, and In the present study, the base sequences in the control
Native American mtDNA on the other, 1,063 contempo- regions of 35 individuals were successfully determined.
rary Native Americans from 16 populations and 36 pre- The sequences in 195 base pairs of HVR 1and 139 base
Hispanic north coast community haplogroup data were pairs of HVR 2 were determined, and mutations were
obtained from published data (Torroni et al., 1993, 1994; observed in 30 portions and 14 portions, respectively
Batista et al., 1995; Santos et al., 1994; Kolman et al., (Table 3). These individuals were classified into 27 hap-
1995; Rickards et al., 1999; Merriwether et al., 1995; lotypes according to their sequences. The rate of successful
Rodriguez-Delfin et al., 2001; Lewis et al., 2005; Baillet recovery and determination of mtDNA sequence varied a
et al., 1994; Ginther et al., 1993; Shimada et al., 2004). good deal from one site to the next. For example, of the
Haplogroup frequencies of these populations were com- 26 individuals from Paucarcancha and 24 from Patallacta,
pared with the ancient individuals under study. 16 and 17 were successfully typed, resulting in high DNA
 TABLE 3. Nucleotide changes observed in ancient Peruvian highlanders analyzed in present study

Site and specimen Maternal Mutations in segments1 APLP analysis3

2
number Haplogroup line 16209–16402 (16000þ) 128–267 10382–10465 (10000þ) 5178 8794 14318 9 bp
Paucarcancha
195 A* A*-1 223 290 319 362 146 235 CRS
T
2
208 A* A*-1 223 290 319 362 146 235 CRS
T
2
216 A* A*-2 217 223 266 290 319 343T 362 146 153 235 260 CRS
x
2
192 B4* B4*-1 217 272 362 CRS CRS


1
213 B4* B4*-2 217 289 143 CRS


1
198 B4* B4*-2 217 289 143 ND


1
203 B4* B4*-3 217 146 215 CRS
x
1
210 B4* B4*-4 217 228 379N 214 CRS


1
212 B4* B4*-5 214 217 262 231N CRS


1
214 B4* B4*-6 217 278 146 215 CRS


1
227 B4* B4*-7 217 357 143 CRS


1
233 B4* B4*-8 217 362 CRS CRS


1
230 B4a B4a-1 217 261 319 CRS CRS


1
193 C* C*-1 223 298 325 327 146 249d 398 400

C 2
204 C* C*-1 223 298 325 327 146 249d 398 400

C 2
211 C* C*-2 223 298 325 327 249d ND

C 2
Patallacta
680 B4* B4*-2 217 289 143 CRS


1
978 B4* B4*-3 217 146 215 CRS


1
681 B4* B4*-9 217 296N 321 363 390 214 234 CRS


1
686 B4* B4*-10 217 152 CRS


1
689 B4* B4*-10 217 152 CRS


1
687 B4* B4*-11 217 CRS CRS


1
974 B4* B4*-11 217 CRS CRS


1
981 B4* B4*-12 217 268 348 378 379 x CRS


1
989 B4* B4*-13 217 294 143 210 CRS


1
677 B4* B4*-14 217 152, 204 CRS


1
683 B4a B4a-2 217 261 CRS CRS


1
976 B4a B4a-3 217 261N 357 143 CRS


1
678 B* B*-1 217 381 CRS 398


1
682 C* C*-1 223 298 325 327 146 195 249d 398 400

C 2
975 C* C*-3 223 246N 298 325 327 373 x 398 400

C 2
676 C* C*-1? 223 298N 325N 327 x 398 400

C 2
977 D* D*-1 325 362N CRS 398 400 A

2
Huata
899 C* C*-1 223 298 325 327 x 398 400

C 2
897 C* C*-4 223 298 325 327 x 392 400

C 2
1
All polymorphic sites are numbered according to revised CRS (Andrews et al., 1999). CRS denotes that sequence of segment is identical to revised CRS, and N indicates not determined.
Suffix T indicates transversion, and d indicates deletion. Deletions are recorded at last possible site. x indicates that PCR product was not obtained at region. Diagnostic polymorphisms
are indicated by bold italic type.
2
Nucleotide change at position 263 in segment 128–267 was observed in all specimens, and therefore is omitted.
3
Diagnostic polymorphisms are emphasized by bold italic type. Dot indicates that no nucleotide change was observed. x indicates that PCR product was not obtained in region.
 ANCIENT DNA ANALYSIS OF PERUVIAN HIGHLANDERS 7
recovery and sequencing rates of 61.5% and 70.8%, re-

Ginther et al. (1993), Horai et al. (1993), Baillet et al. (1994)

spectively. In contrast, of seven individuals from the

Merriwether et al. (1995), Rodriguez-Delfin et al. (1999)

Huata, only two (or 28.6%) were successfully sequenced.
Haplogroup distribution for the total sample was as
follows: 8.6% A, 65.7% B, 22.9% C, and 2.9% D. Hap-

Torroni et al. (1993, 1994), Batista et al. (1995),

logroup frequencies of contemporary Amerindian popula-

Baillet et al. (1994), Merriwether et al. (1995)

Baillet et al. (1994), Merriwether et al. (1995)

tions and ancient north coast samples are also shown in

Santos et al. (1994), Kolman et al. (1995),

Table 4. F-statistics from haplogroup frequencies among
regional populations are shown in Table 5. An exact test
of differentiation between each pair of populations

References
revealed statistically significant differences except be-

TABLE 4. Frequency distribution of mtDNA haplogroups in ancient Peruvian highlanders, ancient north coast community
tween the ancient highlanders and contemporary central
Andean population (significant Fst P ¼ 0.180 6 0.054).
To investigate the relationships among the satellite

Merriwether et al. (1995)

Merriwether et al. (1995)

Rickards et al. (1999)
communities of the royal estate of Machu Picchu, mtDNA

Shimada et al. (2004)

sequences of Paucarcancha and Patallacta were compared.

Lewis et al. (2004)

Haplogroup frequencies of Paucarcancha and Patallacta
are shown in Table 6. Genetic diversity results for these

Present study
two sites are shown in Table 7. Mean numbers of pairwise
differences and nucleotide diversity are slightly larger in
the Paucarcancha.

(Sicán and Sipán), and contemporary Amerindian populations

DISCUSSION

7.1% (26)

1.6% (11)
0% (0)

0% (0)
0% (0)
0% (0)

0% (0)
0% (0)
22.2% (8)

4.3% (6)
4.3% (3)

0.9% (3)

1.7% (6)
Others
Haplogroup profile of individuals examined
in the present study
We found that haplogroup B was the most frequent
among skeletal samples analyzed in the Inca-period resi-

42.2% (151)
28.8% (201)

Numbers in parentheses indicate number of individuals who were assigned to relevant haplogroups.
30.0% (42)
21.4% (15)

14.0% (24)

14.5% (49)

52.0% (52)
48.3% (57)
30.6% (11)
dents of the Urubamba Valley, followed by haplogroups

21.2% (7)
2.9% (1)

2.2% (8)

4 .8% (3)
Haplogroups and their frequencies1

C, A, and finally D. The most distinctive feature of the

D

haplogroup profile of individuals examined in the pres-

ent study is the high frequency of haplogroup B (65.7%;
23 of 35 individuals; Tables 3 and 4). Classifying individ-

27.1% (38)
14.3% (10)

12.2% (21)

12.7% (43)

37.0% (37)
18.6% (22)
27.1% (97)
20.0% (14)
3.0% (11)

uals into maternal lines resulted in haplogroup B having

22.9% (8)

18.2% (6)
5.6% (2)

9.5% (6)
at least 18 different lines in 23 individuals. In other
C

words, the high frequency of haplogroup B is not caused

by the concentration of individuals on a specific maternal
line.
29.5% (108)

63.0% (213)

42.8% (299)
67.4% (116)

Haplogroup B is the common haplogroup in contempo-

65.7% (23)

30.8% (43)
50.0% (35)
51.5% (17)

71.4% (45)

28.8% (34)
24.0% (86)
22.2% (8)

9.0% (9)
rary Central Andean populations. When the haplogroup
B

profile of these ancient residents of the Urubamba Valley

was compared with that of other South American popu-
lations, the former showed a clear proximity to the modern
Central Andean populations that are distributed primarily
58.2% (213)

8.9% (30)

5.0% (18)
6.9% (48)
7.9% (11)
6.4% (11)

in the Peruvian and Bolivian highlands (Table 4). This

19.4% (7)

10.0% (7)

14.3% (9)
8.6% (3)

9.1% (3)

2.0% (2)
4.2% (5)

finding is not surprising, considering the highland location

A

of the study area.

On the other hand, the ancient highlanders consider-
ably differ from individuals of the ancient north coast
Southern Andean (Peheunche)
Central Andean (Atacameños)

community in terms of mtDNA haplogroup frequency.

Contemporary lower Central America (total)

Southern Andean (Huilliche)

Southern Andean (Mapuche)

Various lines of archaeological evidence indicate intimate

Central Andean (Quechua)

Central Andean (Aymara)

Central Andean (Ancash)

cultural interactions between the ancient north coastal

Southern Andean (total)
Central Andean (Total)

populations and contemporaneous Ecuadorian and Co-

lombian populations (Shimada, 1995, 1999; Shimada
Ancient north coast community

et al., 1997, 2000). Relatively high frequencies of hap-

Andean (total)

logroup A and low frequencies of haplogroup C in an-

Population

cient north coast populations hint at their linkage with

modern lower Central American and North Andean pop-
ulations. However, in the cases of samples from ancient
Ancient highlanders

north coast communities and ancient highlanders, there

is the possibility that the frequency for each haplogroup
Contemporary
Contemporary
Contemporary
Contemporary
Contemporary
Contemporary
Contemporary
Contemporary
Contemporary
Contemporary

was biased, because these samples were taken only from

the few sites considered to have been related by blood
relationship. Yet an exact test of differentiation (Ray-
mond and Rousset, 1995) at least shows that the ancient
highlanders bear a similarity to contemporary Central
1
8 K. SHINODA ET AL.
TABLE 5. Pairwise Fst values between each pair of populations
Ancient highlanders Ancient north coast Lower Central America Central Andean
Ancient north coast 0.1956
Lower Central America 0.2633 0.1883
Central Andean 0.0064* 0.1689 0.2528
Southern Andean 0.1869 0.0679 0.2866 0.1648

* Fst P-value between ancient highlanders and contemporary Central Andean population is 0.180 6 0.054 (standard error).

TABLE 6. Frequency distribution of mtDNA haplogroups in ancient Peruvian highlanders excavated from
Paucarcancha and Patallacta sites
Haplogroups and their frequencies1
Population A B C D Others
Paucarcancha 18.8% (3) 62.5% (10) 18.8% (3) 0.0% (0) 0.0% (0)
Patallacta 0.0% (0) 75.0% (12) 18.8% (3) 6.3% (1) 0.0% (0)
1
Numbers in parentheses indicate number of individuals who were assigned to relevant haplogroups.

TABLE 7. mtDNA HV 1 haplotype diversity parameters not discount the distinct possibility that at least some of
of Paucarcancha and Patallacta1 the individuals we analyzed had been relocated by the
n k Pw p Pairwise Fst Inca state to Paucarcancha and Patallacta from other Pe-
ruvian and/or Bolivian highland regions. Furthermore,
Paucarcancha 16 12 4.475 0.023 0.0222 the frequencies of mtDNA haplotypes in the sampled pop-
Patallacta 17 13 3.059 0.016
ulations could be biased due to the small sample size,
1
n, sample size; k, number of different sequences; Pw, mean with the true frequencies likely to be different when more
number of pairwise differences; p nucleotide diversity. individuals are typed. Further genetic studies are needed
to confirm our results.

Andean people on the basis of sharing similar haplo- CONCLUSIONS

groups.
As described above, morphological analysis of the Our mtDNA study showed the clear proximity of an-
Machu Picchu residents suggests that they were a mixed cient residents of the Urubamba Valley near Machu Pic-
group of natives from various coastal and highland chu to modern Quechua and Aymara populations in the
regions. The mixed composition is likely to reflect the Peruvian and Bolivian highlands, respectively. Further,
pervasive Inca practice of recruiting male and female re- these residents were considerably different from pre-
tainers and specialists from subjugated populations and Hispanic individuals of the north coast of Peru in terms
relocating them in accordance with royal and/or state of mtDNA haplogroup frequency. Thus, the results sup-
purposes. Thus, the haplogroup profile of Machu Picchu port the hypothesis that the remains of residents of
individuals is expected to be different from that of resi- the Inca-period rural communities of Paucarcancha and
dents of their rural hinterlands that we examined in the Patallacta that were analyzed were native highlanders,
present study. This question should be solved directly although we cannot define their exact ethnic and geograph-
bythe DNA analysis of skeletons excavated from the ical origins; we argue that they served roles of supporting
Machu Picchu site. the nearby Inca royal estate of Machu Picchu. We suspect
that future ancient DNA analysis of the Machu Picchu
Relationship between sites residents will show that they were genetically distinct
from the sampled individuals in the present study. How-
One of the main purposes of analyzing DNA from an- ever, the fact that mtDNA analysis does not provide any
cient burial sites is to determine whether the human information regarding genetic connections along the
remains represent related or unrelated individuals. As paternal line should be considered. Moreover, character-
described above, our samples were mostly obtained from istics of group genetics are influenced by a number of
two sites (Paucarcancha and Patallacta) that represent factors that include not only natural phenomena, such
satellite support communities of the royal estate of as genetic drift or the bottleneck effect, but also social
Machu Picchu. On comparing the haplotype profiles of and cultural elements, such as marriage, war, or trade.
individuals excavated from the two sites, an interesting Therefore, it is impossible to clarify all the changes that
observation was made. As shown in Tables 6 and 7, both a group had undergone through genetic analysis alone.
populations have high frequencies of haplogroup B, and In order to ascertain the fate of the inhabitants of an
the population pairwise Fst was not so large. Moreover, archaeological site, in addition to clarifying its regional
an exact test of differentiation revealed that these differ- context, genetic data should be viewed as one of multiple
ences are not statistically significant (P ¼ 0.153 6 0.039 independent lines of evidence garnered by a team of
(standard error)). However, as shown in Table 3, at least archaeologists, physical anthropologists, and other com-
13 and 14 maternal lines were distinguished in Paucar- plementary specialists.
cancha and Patallacta, respectively; these two sites
shared only three maternal lines. The result of our ACKNOWLEDGMENTS
mtDNA analysis suggests that the maternal structure of
the Paucarcancha and Patallacta individuals was clearly We are grateful to Elsa Tomasto and Hilda Vidal
different. On the basis of our mtDNA data alone, we can- (curators of the Museo Nacional de Antoropologia
 ANCIENT DNA ANALYSIS OF PERUVIAN HIGHLANDERS 9
Arqueologia e Historia del Peru) and Japanese photogra- Kivisild T, Tolk HV, Parik J, Wang Y, Papiha SS, Bandelt HJ,
pher Yutaka Yoshii for their assistance in the collection Villems R. 2002. The emerging limbs and twigs of the East
of tooth samples used in the mtDNA analysis. Research Asian mtDNA tree. Mol Biol Evol 19:1737–1751.
by K.-I.S. for this study was supported by Grant-in-Aid Kolman CJ, Tuross N. 2000. Ancient DNA analysis of human
populations. Am J Phys Anthropol 111:5–23.
for Scientific Research 13575017 from the Ministry of
Kolman CJ, Bermingham E, Cooke R, Ward RH, Arias TD,
Education, Science, Sports and Culture, Japan. Guionneau-Sinclair F. 1995. Reduced mtDNA diversity in the
Ngobe Amerinds of Panama. Genetics 140:275–283.
Kong QP, Yao YG, Sun C, Bandelt HJ, Zhu CL, Zhang YP. 2003.
Phylogeny of East Asian mitochondrial DNA lineages inferred
LITERATURE CITED from complete sequences. Am J Hum Genet 73:671–676.
Adachi N, Dodo Y, Ohshima N, Doi N, Yoneda M, Matsumura Lewis CM, Tito RY, Lizárraga B, Stone AC. 2005. Land, lan-
H. 2003. Morphologic and genetic evidence for the kinship of guage, and loci: mtDNA in Native Americans and the genetic
juvenile skeletal specimens from a 2,000 year-old double bur- history of Peru. Am J Phys Anthropol 127:351–360.
ial of the Usu-Moshiri site, Hokkaido, Japan. Anthropol Sci MacCurdy V, Richards M, Hickey E, Vega E, Cruciani F, Guida
111:347–363. V, Scozzari R, Bonné-Tamir, B, Sykes B, Torroni A. 1999. The
Adachi N, Umetsu K, Takigawa W, Sakaue K. 2004. Phyloge- emerging tree of West Eurasian mtDNAs: a synthesis of con-
netic analysis of the human ancient mitochondrial DNA. J trol-region sequence and RFLPs. Am J Hum Genet 64:232–249.
Archaeol Sci 31:1339–1348. MacCurdy G. 1918. Surgery among ancient Peruvians. Art
Alves-Silva J, Santos MDS, Guimarães PEM, Ferreira ACS, Archaeol 7:381–394.
Bandelt HJ, Pena SDJ, Prado VF. 2000. The ancestry of MacCurdy G. 1923. Human skeletal remains in the Peruvian
Brazilian mtDNA linages. Am J Hum Genet 67:444–461. highlands. Am J Phys Anthropol 6:217–330.
Andrews RM, Kubacka I, Chinnery PF, Lightowlers RN, Turn- Maruyama S, Minaguchi K, Saitou N. 2003. Sequence polymor-
bull DM, Howell N. 1999. Reanalysis and revision of the Cam- phisms of the mitochondrial DNA control region and phyloge-
bridge reference sequence for human mitochondrial DNA. Nat netic analysis of mtDNA lineages in the Japanese popula-
Genet 23:147. tions. Int J Legal Med 117:218–225.
Bailliet G, Rothhammer F, Carnese FR, Bravi CM, Bianchi NO. Merriwether DA, Rothhammer F, Ferrell RE. 1995. Distribution
1994. Founder mitochondrial haplotypes in Amerindian popu- of the four founding lineage haplotypes in Native Americans
lations. Am J Hum Genet 55:27–33. suggests a single wave of migration for the New World. Am J
Bandelt HJ, Lahermo P, Richards M, Macaulay V. 2001. Detect- Phys Anthropol 98:411–30.
ing errors in mtDNA data by phylogenetic analysis. Int J Nata M, Adachi N, Hashiyada M, Kanetake J, Ji G, Mimasaka
Legal Med 115:64–69. S, Takahashi K, Sagisaka K, Funayama M. 1999. Sequence
Batista O, Kolman CJ, Bermingham E. 1995. Mitochondrial polymorphism of the D-loop of mitochondrial DNA in Japa-
DNA diversity in the Kuna Amerinds of Panama. Hum Mol nese and Chinese population. Acta Crim Japan 65:244–254.
Genet 4:921–929. Niles SA. 1984. Architectural form and social function in Inca
Bingham H. 1912. Preliminary report of the Yale Peruvian expe- towns near Cuzco. In: Kendall A, editor. Current archaeologi-
dition. Bull Am Geogr Soc 44:20–26. cal projects in the Central Andes—some approaches and
Bingham H. 1913. In the wonderland of Peru. Natl Geogr Mag results.British Archaeological Reports international series
24:387–573. 210. Oxford: British Archaeological Reports. p 205–219.
Bingham H. 1916. Further explorations in the land of the Incas. Niles SA. 1988. Looking for ‘‘lost’’ Inca palaces. Expedition
Natl Geogr Mag 29:431–473. 30:56–64.
Bingham H. 1930. Machu Picchu, a citadel of the Incas. Report Niles SA. 1999. Shape of Inca history: narrative and architecture
of the explorations and excavations made in 1911, 1912 and in an Andean empire. Iowa City: University of Iowa Press.
1915 under the auspices of Yale University and the National Protzen JP. 1993. Inca architecture and construction at Ollan-
Geographic Society. New Haven: Yale University Press. taytambo. Oxford: Oxford University Press.
Burger RL, Salazar LC. 2003. Preface. In: Burger R, Salazar Quintana-Murci L, Semino O, Bandelt HJ, Passarino G, McEl-
LC, editors. The 1912 Yale Peruvian scientific expedition col- reavey K, Santachiara-Benerecetti AS. 1999. Genetic evidence
lection from Machu Picchu: human and animal remains, Yale for an early exit of Homo sapiens sapiens from Africa through
University publications in anthropology 85. New Haven: eastern Africa. Nat Genet 23:437–441.
Peabody Museum of Natural History, Yale University. p xiii–xv. Raymond M, Rousset F. 1995. An exact test of population differ-
Eaton GF. 1916. The collection of osteological material from entiation. Evolution 49:1280–1283.
Machu Picchu. Memoirs of the Connecticut Academy of Arts Rickards O, Martinez-Labarga C, Lum JK, De Stefano GF,
and Sciences 5. New Haven: Tuttle, Morehouse and Taylor. Cann RL. 1999. mtDNA history of the Cayapa Amerinds of
Farrington IS. 1995. The mummy, estate and palace of Inka Ecuador: detection of additional founding lineages for the
Huayna Capac at Quispeguanca. Tawantinsuyu 1:55–65. Native American populations. Am J Hum Genet 65:519–530.
Ginther C, Corach D, Penacino GA, Rey JA, Carnese FR, Hutz Rodriguez-Delfin LA, Rubin-de-Celis VE, Zago MA. 1999.
MH, Anderson A, Just J, Salzano FM, King MC. 1993. Genetic diversity in an Andean population from Peru and re-
Genetic variation among the Mapuche Indians from the Pata- gional migration patterns of Amerindians in South America:
gonian region of Argentina: mitochondrial DNA sequence data from Y chromosome and mitochondrial DNA. Hum Hered
variation and allele frequencies of several nuclear genes. EXS 51:97–106.
67:211–219. Rowe JH. 1982. Policies and institutions relating to the unifica-
Horai S, Kondo R, Nakagawa-Hattori Y, Hayashi S, Sonoda S, tion of the empire. In: Collier GA, Rosaldo RI, Wirth JD, edi-
Tajima K. 1993. Peopling of the Americas, founded by four tors. The Inca and Aztec states, 1400–1800. New York: Aca-
major lineages of mitochondrial DNA. Mol Biol Evol 10:23–47. demic Press. p 93–118.
Kendall A. 1974. Architecture and planning at the Inca sites in Rowe JH. 1990. Machu Picchu a la luz de documentos del siglo
the Cusichaca area. Baessler Archiv 22:73–137. XVI. Historica 14:139–154.
Kendall A. 1985. Aspects of Inca architecture: description, func- Santos M, Ward RH, Barrantes R. 1994. mtDNA variation in
tion and chronology. BAR international series, 242. Oxford, Brit- the Chibcha Amerindian Huetar from Costa Rica. Hum Biol
ish Archaeological Reports. 66:963–77.
Kendall A. 1988. Inca planning north of Cuzco between Anta Schneider S, Roessli D, Excoffier L. 2000. Arlequuin version 2.000:
and Machu Picchu and along the Urubamba Valley. In: a software for population genetics data analysis. Geneva: Genetic
Saunders NJ, de Montomollin O, editors. Recent studies in and Biometry Laboratory, University of Geneva.
pre-Columbian archaeology. Oxford: British Archaeological Shimada I. 1995. Cultura Sicán: dios, riqueza y poder en la
Reports. p 457–488. costa norte del Peru. Lima: Banco Continental.
10 K. SHINODA ET AL.
Shimada I. 1999. The evolution of Andean diversity: regional Y-chromosome polymorphisms in four Native American popu-
formations, ca. 500 B.C.–A.D. 600. In: Salomon F, Schwartz S, lations from southern Mexico. Am J Hum Genet 54:303–318.
editors. Cambridge history of native peoples of the Americas. Umetsu K, Tanaka M, Yuasa I, Saitou N, Takeyasu T, Fuku N,
Cambridge: Cambridge University Press. p 350–517. Naito E, Ago K, Nakayashiki N, Miyoshi A, Kashimura S,
Shimada I, Anderson KB, Haas H, Langenheim JH. 1997. Watanabe G, Osawa M. 2001. Multiplex amplified product-
Amber from 1000-year old prehispanic tombs in northern length polymorphism analysis for rapid detection of human
Peru. In: Vandiver PB, Druzik JR, Merkel JF, Stewart J, edi- mitochondrial DNA variations. Electrophoresis 22:3533–3538.
tors. Materials issues in art and archaeology V. Pittsburgh: Verano JW. 2003. Human skeletal remains from Machu Picchu:
Materials Research Society. p 3–18. a reexamination of the Yale Peabody Museum’s collections. In:
Shimada I, Gordus A, Jo A. 2000. Technology, iconography, and Burger R, Salazar L, editors. The 1912 Yale Peruvian scien-
significance of metals: a multi-dimensional analysis of Middle tific expedition collection from Machu Picchu: human and ani-
Sicán Objects. In: McEwan C, editor. Pre-Columbian gold: mal remains. Yale University publications in anthropolology
technology, iconography, and style. London: British Museum 85. New Haven: Peabody Museum of Natural History, Yale
Press. p 28–61. University. p 65–117.
Shimada I, Shinoda K, Farnum J, Corruccini R, Watanabe H.
Von Hagen A, Morris C. 1998. The cities of the ancient Andes.
2004. An integrated analysis of pre-Hispanic mortuary prac-
London: Thames and Hudson.
tices: a Middle Sican case study. Curr Anthropol 45:369–402.
Thomas M, Gilbert P, Willerslev E, Hansen AJ, Barnes I, Rud- Wallace DC, Garrison K, Knowler WC. 1985. Dramatic founder
beck L, Lynnerup N, Cooper A. 2003. Distribution patterns of effects in Amerindian mitochondrial DNAs. Am J Phys
postmortem damage in human mitochondrial DNA. Am J Anthropol 68:149–155.
Hum Genet 72:32–47. Yao YG, Kong QP, Bandelt HJ, Kivisild T, Zhang YP. 2002. Phy-
Torroni A, Schurr TG, Cabell MF, Brown MD, Neel JV, Larsen logenetic differentiation of mitochondrial DNA in Han Chi-
M, Smith DG, Vullo CM, Wallace DC. 1993. Asian affinities nese. Am J Hum Genet 70:635–651.
and continental radiation of the four founding Native Ameri- Yao YG, Kong QP, Man XY, Bandelt HJ, Zhang YP. 2003. Recon-
can mtDNAs. Am J Hum Genet 53:563–590. structing the evolutionary history of China: a caveat about
Torroni A, Chen YS, Semino O, Santachiara-Beneceretti AS, inferences drawn from ancient DNA. Mol Biol Evol 20:214–
Scott CR, Lott MT, Winter M, Wallace DC. 1994. mtDNA and 219.