Introduction

The Caucasus is a region located in the border between Europe and western Asia (between the Black and Caspian Seas). This region exhibits a high degree of linguistic diversity, with four major language families represented (North Caucasian, South Caucasian, Indo-European, and Altaic), which are spoken by around 50 autochthonous groups. A notable geographic barrier, the Caucasus Mountains, divides the northern and southern sub-regions of the Caucasus. Georgians are the main ethnic group in Swanetia, a province of the Republic of Georgia on the southern slopes of the Central Caucasus. Its rough, mountainous terrain has helped make Swanetia a very isolated region, and this has been decisive in the preservation of its cultural and linguistic patrimony, and therefore of its genetic background. The local language (Svan) belongs to the Kartvelian or South Caucasian language family and is non Indo-European in origin (Gamkrelidze and Ivanov 1990; Renfrew 1991).

The D-loop region of human mitochondrial DNA (mtDNA), which is approximately 1,150 base pairs (bp) long, is located between the mitochondrial tRNAPro and tRNAPhe genes. This is the most variable region in the mitochondrial genome, and the most polymorphic nucleotide sites within this loop are concentrated in two ‘hypervariable segments’, HVI (positions 16024–16365) and HVII (positions 69–340) (Aquadro and Greenberg 1983).

Mitochondrial DNA is maternally inherited and evolves rapidly. The average number of base pair differences between two human mitochondrial genomes is estimated to be between 9.5 and 66 (Zeviani et al. 1998). The high mutation rate has resulted in the accumulation of a wide range of neutral, population-specific base substitutions in mtDNA. These have accumulated sequentially along radiating maternal lineages that have diverged approximately on the same time-scale as human populations have colonized different geographical regions of the world (Torroni and Wallace 1994; Wallace 1995). In the last two decades, genetic data derived from mtDNA studies have helped greatly in elucidating human evolution, estimating a time-scale for events in human prehistory, and detecting past demographic movements (Vigilant et al. 1991; Schurr and Wallace 2002; Tajima et al. 2004; Starikovskaya et al. 2005). The study of female lineages has provided authentic evidence on the African origin and subsequent dispersal of our species (Cann et al. 1987).

On the other hand, dysfunction in mitochondrial processes has been related to several pathologies. Mutations in mtDNA are now recognized as major contributors to human pathologies and possibly to normal aging (reviewed in Wallace 1997, 1999; Zeviani et al. 1998). A large number of rearrangements and point mutations in protein-coding and tRNA genes have been identified in patients with mitochondrial disorders (Holt et al. 1988; Reardon et al. 1992; Procaccio and Wallace 2004; Sudo et al. 2004).

Finally, analysis of mtDNA is a very useful forensic tool (Parson et al. 1998). When forensic cases arise where there is insufficient biological material for nuclear DNA typing, mtDNA analysis can provide valuable supplemental information, even from such limited samples as 0.5-cm-long hair fragments or single teeth. Because of its usefulness when limited biological material is available, and due to its unique pattern of maternal inheritance, mtDNA plays a key role in personal identity testing (Budowle et al. 2003).

In recent years, many databases of the mtDNA control region have been published to permit mtDNA forensic casework in particular countries or geographic regions (Parson et al. 1998; Rousselet and Mangin 1998; Budowle et al. 1999, 2002; Pfeiffer et al. 1999; Imaizumi et al. 2002; Vanecek et al. 2004; Zupanic Pajnic et al. 2004). However, the bulk of mtDNA sequence data published to date have been centered in the HVI segment. The strength of the mtDNA evidence that can be reported is very often limited by the lack of relevant database information. For this reason, it is important that mtDNA sequence databases continue to be generated to extend mtDNA typing capability to additional populations and to increase the size of existing databases (Imaizumi et al. 2002).

The genetic characterization of so-called ‘isolated’ human populations is relevant to accomplishing several scientific goals such as those related to disease mapping, human demographic history and forensic identification (Arcos-Burgos and Muenke 2002; Pérez-Miranda et al. 2004, 2005). Previous studies addressing the genetic characterization of mtDNA polymorphism in Georgians have been limited to the HVI segment (Comas et al. 2000; Nasidze and Stoneking 2001). The aim of the present study was to analyze HVI and HVII sequences of the mtDNA control region in a sample of autochthonous individuals from Swanetia (Georgia) using direct fluorescent-based sequencing. Our findings on mtDNA diversity in Georgia are then viewed in a comparative context, using previously published data on mtDNA haplogroups for other human populations worldwide. With this integrative approach, we seek to assess population affinities and phylogenetic relationships of the Georgian group in a broader geographical context. The genetic information presented herein may be highly useful in both evolutionary genetic studies and forensic analyzes, especially data on the polymorphism of hypervariable region HVII, which is analyzed here for the first time in Georgians.

Materials and methods

Genomic DNA was extracted from peripheral blood using a standard phenol-chloroform procedure (Sambrook et al. 1989). Blood samples were collected by venipuncture from 48 unrelated healthy men living in the Caucasus region (Georgia). Only native people were included in the sample; Georgian ancestry (based on surnames and birthplaces) was ascertained for three generations back in order to define autochthony for each donor. The autochthonous individuals were carefully selected to avoid the bias caused by sample heterogeneity. Specifically, the Georgian population sample was collected in the region of Swanetia to avoid any admixture with Turkish and Russian lineages, which are most frequently found in the lowlands of the country. Ethical guidelines were adhered to as stipulated by each of the institutions involved in the study. All blood donors gave their informed consent prior to their inclusion in the study sample.

The HVI and HVII segments of the mtDNA D-loop region were amplified by polymerase chain reaction (PCR). The primers used to amplify HVI and HVII segments were L15996/H16401 and L29/H408, respectively, as described by Vigilant et al. (1991). PCRs were performed in a Thermal Cycler GeneAmp PCR System 9600 (PE Applied Biosystems, Foster City, CA) by hot start at 85°C and 30 cycles of 94°C 45 s, 66°C 60 s and 72°C 60 s, and a final extension step at 72°C for 10 min. PCR products were sequenced with the dRhodamine Terminator Cycle Sequencing Ready Reaction kit on an ABI Prism 310 Genetic Analyzer (PE Applied Biosystems).

The mtDNA sequences reported in this paper have been deposited in the DDBJ/EMBL/GenBank nucleotide database (accession numbers AM181177–AM181225).

Statistical analysis

The individual sequences were aligned and compared to the revised Cambridge Reference Sequence for human mtDNA, rCRS (Andrews et al. 1999), by means of the Clustal W program package (Version 1.5). The rCRS is a modified version of the original CRS of Anderson et al. (1981), and has 18 annotated nucleotides. We chose the Multiple Sequence Alignment option (Higgins et al. 1992) to determine the specific nucleotide substitutions. The HVI segment was examined for positions 16033–16362 (according to rCRS), while analysis of the HVII segment included positions 66–370. The number of constant sites was calculated using Puzzle (Version 4.02) software (Strimmer and Von Haeseler 1995). The genetic variability of the mtDNA sequences in the Georgian population was measured by means of three parameters: nucleotide diversity (Nd), sequence diversity, and haplogroup diversity (Nei 1987). These basic diversity parameters were computed using the Arlequin program, version 2.000 (Schneider et al. 2000). The random match probability (RMP) was estimated as the solution to the equation

$$ RMP = {\sum\limits_{i = 1}^h {p^{2}_{I} } }, $$

where pI is the sample frequency of the I-th haplotype, and h is the number of distinct haplotypes observed in the sample (Jones 1972).

For interpopulation comparison purposes, published data on both HVI haplogroups and sequence variations in the hypervariable regions HVI and HVII from many other human groups worldwide were compiled. Thus, mtDNA haplogroup data in Georgians were compared to the following 23 populations: Andalusian, Catalonian, Portuguese, Algerian (Côrte-Real et al. 1996), British (Piercy et al. 1993), North Portuguese (Pereira et al. 1999), Tuscan (Francalacci et al. 1996), South German (Lutz et al. 1998), North-West German (Pfeiffer et al. 1999), Austrian, Spanish, Egyptian, Japanese, sub-Saharan African, African-Americans (USA), European Americans (USA), Hispanic, Navajo (Budowle et al. 1999), Basque (unpublished data), Kung (Vigilant et al. 1991), French (pooled from Rousselet and Mangin 1998; Budowle et al. 1999), Korean (pooled from Pfeiffer et al. 1998; Budowle et al. 1999), and Taiwanese (pooled from Melton et al. 1995; Budowle et al. 1999).

Genetic distances between pairs of populations were calculated using the mismatch-intermatch means of the pairwise difference distribution. Then, D=dij−(dii+djj)/2, where dij is the intermatch mean (i.e., the mean number of nucleotide differences) between populations I and J, and dii and djj are the mismatch values within populations I and J, respectively (Rao 1982; Nei 1987). In order to depict the mismatch–intermatch distance matrix in a two-dimensional genetic map, nonmetric multidimensional scaling (MDS) analysis (Kruskal 1964) was performed using the SPSS (version 13.0; SPSS, Chicago, IL) statistical package. Furthermore, phylogenetic trees based on the neighbor-joining (NJ) method (Saitou and Nei 1987) were generated from the resultant distance matrix, using programs in Phylip version 3.2 (Felsenstein 1989). The reliability of the dendrogram (robustness of the branch nodes) was estimated by bootstrap resampling methods (Felsenstein 1985). Finally, population trees based solely on the sequence variations for HVI, on the one hand, and on the sequence variations identified when a joint analysis of HVI + HVII was carried out on the other hand, were compared. In this latter analysis, only 14 populations (including Georgia) could be considered, due to the paucity of HVII data. Populations included were those compiled in Budowle et al. (1999) and Basques (unpublished data).

Results and discussion

Sequence variability

Sequencing data on the hypervariable region HVI, corresponding to nucleotide positions 16033–16362 in the revised Cambridge Reference Sequence for human mtDNA (rCRS), are shown in Table 1. Sequence comparison between the Georgian individuals (n=48) and the rCRS led to the identification of 37 different mtDNA haplotypes in HVI, defined by 46 polymorphic positions. Of them, 28 mtDNA types (75.7%) were unique, seven haplotypes were found twice, and two haplotypes were shared by three individuals. Only one mtDNA sequence showed no difference with regard to the rCRS (code GS-78, in Table 1).

Table 1 Mitochondrial DNA diversity of the hypervariable region HVI in a Georgian population (n=48). Polymorphic sites are given in comparison to the revised Cambridge reference sequence (Andrews et al. 1999), rCRS (bold type at the top)
Full size table

The number of constant positions in HVI was 284, which represents 86.1% of the mtDNA fragment analyzed (330 bp in length). Most of the variable positions identified in the HVI segment were transitions, specifically 38, which is equivalent to 82.6% of the polymorphic sites. Furthermore, seven transversions and one insertion were also detected (see also Table 3).

The most frequent nucleotide substitutions were the transitions 16311 T>C in 16 individuals (33.3%) and 16223 C>T in 10 individuals (20.8%). Other common transitions observed were 16224 T>C and 16294 C>T, which appeared in eight individuals each (16.7%), whereas transitions 16126 T>C and 16189 T>C were both found in 14.6% of cases (seven individuals). Likewise, the most frequent transversion in the Georgian sample analyzed was 16183 A>C, which was identified in four individuals. Other transversions detected were 16085 C>G, 16146 A>C, 16071 A>C, 16214 C>A, 16246 A>T, and 16359 T>G, each of them observed in one individual. In addition, an insertion in the region delimited by 16190–16193 bp (four individuals) was found, denoted as 16193.1C. There was length heteroplasmy in the homopolymeric C-stretch region (located between nucleotide positions 16184 and 16193 of HVI) in 10.4% of the sequences examined.

Analysis of mtDNA hypervariable region HVII included 305 bp (positions 66–370, Table 2). Sequence comparisons with rCRS permitted the identification of 31 different mtDNA variants defined by 32 polymorphic sites. Of the whole set of mitochondrial haplotypes, 22 DNA sequences (71.0%) were unique, five haplotypes were shared by three individuals, three haplotypes were found twice, and the most common haplotype was identified in five individuals.

Table 2 Mitochondrial DNA diversity of the hypervariable region HVII in a Georgian population (n=48). Polymorphic sites are given in comparison to the rCRS (Andrews et al. 1999); bold type at the top
Full size table

A more in-depth analysis of mtDNA hypervariable region HVII revealed 273 constant positions (89.5%) in the sequences examined. Of the total number of nucleotide changes detected in this study, 23 (71.9%) were transitions, 4 (12.5%) were transversions and 5 (15.6%) were insertions (see also Table 3). Interestingly, the transition 263 A>G was present in all the sequences analyzed (100%). Apart from this relevant nucleotidic change, the most frequent transitions in the HVII segment were 73 A>G and 195 T>C, which were found in 42 (87.5%) and 21 (43.8%) individuals, respectively. The four transversions observed in HVII were located at nucleotide positions 186 C>A, 194 C>A, 295 C>A (in two individuals) and 298 C>A.

Table 3 Sequence polymorphism of mitochondrial DNA hypervariable regions HVI and HVII in a sample of a Georgian population. CP Constant positions, Ps polymorphic sites, Ts transitions, Tv transversions, Ins insertions, Nd nucleotide diversity, Sd sequence diversity, MNM mean number of mismatches, RMP random match probability
Full size table

There is also a poly-C tract between positions 303 and 315 in the HVII segment, which is interrupted at position 310 by a thymine according to the rCRS. In this study, we identified some sequences that presented between 1 and 3 insertions in the region of nucleotides 303–309. Specifically, we observed 32 cases (66.6%) showing one insertion (309.1C). A second C insertion (309.2C) was detected in three individuals (6.3%), whereas a third insertion (309.3C) was found in two sequences (4.2%). On the other hand, insertion 315.1C was present in all individuals (100%) in the sample. Length heteroplasmy of the homopolymeric poly-C tract was observed in five sequences (10.4%). It is worth highlighting that the most frequent haplotypes in the Georgian population, i.e., 263 (G), 309.1 (C) and 315.1 (C) have been also reported as the most frequent in other European populations. This fact supports the hypothesis that these haplotypes might represent an ancient European common mtDNA sequence (Zupanic Pajnic et al. 2004).

The joint analysis of both segments (HVI and HVII) yielded 43 different haplotypic combinations. The number of constant positions was 557 (87.7%), whereas the number of polymorphic sites was 78 (see Table 3), with wide predominance of transitions (78.2%) over transversions (14.1%) and insertions (7.7%).

Diversity parameters

Some parameters characterizing within-population diversity of the mtDNA sequence, such as nucleotide diversity (Nd) and sequence diversity (Sd), are listed in Table 3. The Nd was higher in the hypervariable region HVI (0.017) than in HVII (0.011). The Nd estimated for HVI was slightly above the top edge of the range (0.0081–0.0155) reported in previous studies including several European populations (Salas et al. 1998; Zupanic Pajnic et al. 2004). This diversity parameter showed values similar to those observed in West Asian samples and in populations from the Anatolian Peninsula (e.g., Turkey), where figures of around 0.016 have been reported (Mergen et al. 2004). Interestingly, the Nd computed for HVII falls far below the variation range (0.0137–0.0187) reported for European populations in recent works (Salas et al. 2000; Zupanic Pajnic et al. 2004).

Consistent with the results described above for Nd, Sd was higher in the HVI region (0.985) than in HVII (0.975). The Sd value calculated for HVI is comparatively higher than levels reported in previous works for other European populations such as British (0.973) (Piercy et al. 1993) and Spaniards (0.939) (Côrte-Real et al. 1996). The Sd value obtained for the population examined (Georgians from Swanetia) was also slightly higher than that estimated in samples from other regions of Georgia, where values between 0.964 and 0.971 have been found (Comas et al. 2000; Nasidze and Stoneking 2001). Sequence diversity in Georgians was, however, very similar to that obtained in populations located in the Anatolian Peninsula, such as Turks (0.988) and Kurds (0.985) (Comas et al. 2000).

As expected, the Sd value increased up to 0.994 when both hypervariable regions (HVI and HVII) were considered jointly. This figure falls into the Sd range estimated by Budowle et al. (1999) for several worldwide populations (0.990–0.998), although it remains below those calculated in several European populations such as those of Austrian, French and USA-Europid samples (0.996). It must be emphasized that Sd was relatively high in Georgians solely in the case of the HVI segment. When we computed Sd including HVII, the global value (HVI+HVII) of Sd was comparatively lower than the estimated Sd in many other European populations analyzed to date (see Budowle et al. 1999; Zupanic Pajnic et al. 2004). These results are in agreement with previously discussed results regarding Nd.

Finally, we estimated the probability of two randomly selected sequences matching (RMP), which is inversely proportional to the power of discrimination of a given genetic marker. The RMP value was lower in hypervariable region HVI (3.4%) than in HVII (4.5%) (Table 3). In a logical reflection of the data on diversity parameters (Nd and Sd), the RMP value estimated for the Georgian population considering HVI and HVII (2.6%) was clearly above the variation range estimated for the Spanish regions of Aragon (1.3%) and Madrid (1.5%), and for other European populations such as those of Switzerland (0.9%), France (2.1%), Germany (0.9%), Austria and the United Kingdom (1.3%) (Dimo-Simonin et al. 2000). These findings suggest that the variability of maternal lineages tends to be lower in traditional human isolates, most likely because of extremely restricted gene flow and high inbreeding levels over long periods. Thus, the power of discrimination of the mtDNA in forensic identity testing is thought to be more limited in this type of population. This hypothesis is clearly fulfilled in another interesting ‘human isolate’ from Europe: the Basques (work in progress). The RMP estimated for HVI in the native Basque population stands out as having a notably high value (10.1%). The RMP for HVII, although lower than that of HVI, is also comparatively high (4.7%). Likewise, the RMP value calculated taking both HVI and HVII into account is practically twice as high (1.96:1.00) in Basques (5.1%) than in Georgians (2.6%), and these differences are even greater when compared to other Europeans.

MtDNA haplogroups and genetic relationships with other populations

MtDNA haplogroup-based analyses are very important in unveiling genetic relationships among different human groups. Haplogroup frequencies in the Georgian population are depicted in Fig. 1. MtDNA haplogroups were inferred from control region sequences (HVI+position 73 of HVII) according to the classification proposed by Richards et al. (1998). Within the set of mtDNA haplogroups identified in the study sample, only four showed frequencies above 10%. These were haplogroups U (25.0%), T (10.4%), K and H (12.5% each). As can be noted, haplogroup U was the more frequent in Georgians, in contrast with most European populations analyzed to date, where haplogroup H stands out as the most abundant. Another peculiarity of the sample analyzed herein is the absence of haplogroup U5. This haplogroup was identified in the Georgian collection examined by Comas et al. (2000). The haplogroup diversity estimate yielded a value of Hd=0.891.

Fig. 1
figure 1

Haplogroup frequencies of mitochondrial DNA (mtDNA) in a Georgian population (Swanetia). Haplogroups were inferred from control region sequences (HVI+position 73 of HVII) according to the classification proposed by Richards et al. (1998). Frequency values (%) are displayed. Haplogroups not represented in the histogram were not found

Full size image

In order to assess the genetic relationships of Georgians with other populations based on mtDNA diversity, we compiled haplogroup data from previously published papers (see Materials and methods). First, we computed the mismatch-intermatch distances (Rao 1982; Nei 1987) between all pairs of populations (data not shown). A nonmetric MDS analysis was then used to represent the data generated by the distance matrix in a two-dimensional space.

Figure 2 illustrates the MDS plot generated from the distance matrix. Populations were grouped according to geography, and this topology was highly robust from the statistical viewpoint, bearing in mind that the two-dimensional representation accounted for 98.6% of the total variance. A major division can be observed along Dimension I between the Europid group (European and North African populations), which concentrated in the negative segment, and the Asian and sub-Saharan African samples (grouped in the positive portion of Dimension I). North African populations (Algeria, Egypt) occupied the most remote positions within the Europid cluster. Overall, the samples related to the African continent (Algerians, Egyptians, sub-Saharan Africans, Afro-Americans and Kung) tended to concentrate at the negative edge of Dimension II. Of these populations, the Kung sample seems to be the most differentiated genetically, as can be inferred from its remote position from the centroid of the distribution.

Fig. 2
figure 2

Nonmetric multidimensional scaling (MDS) illustrating the genetic relationships derived from mtDNA haplogroup (according to Richards et al. 1998) data among Georgians and 23 other worldwide populations. The MDS plot was constructed based on the mismatch–intermatch means of the pairwise difference distribution (Rao 1982; Nei 1987). The total variance accounted for in the two-dimensional representation is 98.6%. Geor Georgia, Andl Andalusia, Ctln Catalonia, Basq Basques, Span Spain, NPort North Portugal, Port Portugal, Tusc Tuscany, Fran France, SGerm South Germany, NWGerm North-West Germany, Aust Austria, Brit British, US-Eur United States’ Europids, Egyp Egypt, Algr Algeria, Afri sub-Saharan African, US-Afri Afro-Americans, Kung Kung, Tawn Taiwan, Japn Japan, Kore Korea, Hisp Hispanics, Navj Navajo

Full size image

Worthy of attention is also the central position of the Georgian sample, which plotted close to the centroid, in a practically intermediate position between Europeans and Asians. Two plausible scenarios have been used in some previous studies to explain this interesting pattern: one scenario corresponds to a strictly phylogenetic interpretation of population trees and topogenetic maps, while other interprets these data strictly in terms of migration. The first scenario points out that Caucasus groups are derived from Near Eastern groups and are immediately ancestral to Europeans (Nasidze and Stoneking 2001). The second scenario argues that Caucasus groups are admixed and have experienced gene flow from both Europe and Asia. In this respect, although findings of some previous investigations have revealed a low degree of genetic divergence among Europid populations according to levels of differentiation observed in the sequence of the mtDNA control region, some authors indicate that there is a certain geographic patterning of the mtDNA variation, probably denoting a stepping-stone position of the Anatolian Peninsula (Asian Turkey) and the Caucasus region (including Georgia) between the Middle East and Europe (Calafell et al. 1996; Comas et al. 1996). This explanation is based on the hypothesis of the replacement of Neanderthals in Europe by the arrival of anatomically modern humans during the cultural expansion of the Upper Paleolithic age, between 50,000 and 100,000 years ago.

To verify whether information on HVII sequence polymorphism could contribute to improving a phylogenetic classification based on mtDNA data, population trees were generated from the mismatch-intermatch distance matrix using the NJ method (Fig. 3). NJ trees were obtained using two different data sets: (1) mtDNA sequence variation solely in the hypervariable region HVI (Fig. 3a), and (2) mtDNA sequence variation for hypervariable regions HVI+HVII considered jointly (Fig. 3b). The thorough comparison of the population trees reveals an improvement in the phylogenetic classification when data on HVI and HVII are considered simultaneously. This improvement is reflected both in a substantial increase of the average bootstrap value and in a more adjusted distribution of the human groups, as expected bearing in mind the geographical and ethnohistorical characteristics of the populations included. Thus, the average bootstrap value of the NJ trees increased from 41.9 (HVI) up to 52.8 (approximately 1.3 times higher) when HVII data were incorporated into the analysis. Furthermore, the conspicuously doubtful locations of Austria (in an African cluster) on the one hand, and of Japan—separated from the rest of Asian populations and showing the greatest genetic affinity with Hispanics (see Fig. 3a)—were remarkably improved by including data on HVII polymorphic sites (see Fig. 3b). Consistent with the findings of the MDS, the Georgian population segregated alone, in an intermediate position among the main population clusters. This result reinforces the idea that the Caucasus region might have functioned as a migratory corridor for human displacements between Africa, Asia and Europe.

Fig. 3a,b
figure 3

Neighbor-joining (NJ) trees based on mismatch-intermatch distance (Rao 1982; Nei 1987) applied on mtDNA data for 14 populations worldwide. Figures in tree nodes are percentage bootstrap values, estimated from 1,000 iterations. a NJ tree obtained using data on sequence variation for HVI. b NJ tree obtained using data on sequence variation for HVI + HVII. Geor Georgia, Basq Basques, Span Spain, Fran France, Aust Austria, US-Eur United States’ Europids, Egyp Egypt, Afri sub-Saharan African, US-Afri Afro-Americans, Tawn Taiwan, Japn Japan, Kore Korea, Hisp Hispanics, Navj Navajo

Full size image

Both the MDS data and the NJ data failed to identify the remnants of a putative Paleolithic background shared between Georgians and Basques, as some authors have claimed based mainly on linguistic criteria (Lafon 1951). If the Caucasian and Basque languages really are related pre-Indo-European languages of Paleolithic antiquity, one might expect to see evidence of genetic affinity between both groups. Nevertheless, neither phylogenetic results (MDS and NJ data) derived from this study nor the forensic data (RMP) support such a controversial hypothesis. The genetic heterogeneity observed between Georgians and Basques with respect to mtDNA could be generated by differences in mutation rates, in demographic histories and differences regarding impact of the genetic drift. This situation would in turn be reinforced by the effects of a limited gene flow resulting from isolation by distance, all of which may have obscured any common mtDNA background between both human groups (Bertorelle et al. 1995; Comas et al. 2000).

To summarize, the Sd increased markedly when both mtDNA segments were analyzed jointly (HVI: 0.985, HVII: 0.975, HVI+VII: 0.994) and, consequently, the RMP decreased from 0.034 (HVI) to 0.026 (HVI+HVII). In other words, incorporation of the HVII segment into the analysis increases the power of discrimination of the mtDNA. Similarly, incorporation of HVII data contributed to the refinement of results regarding the genetic relationships and population affinities among the samples included in the analyses. These results unquestionably stress the importance of considering HVII in both population and forensic genetics.