Salta is a province in northwest Argentina. Before the Spanish conquest, numerous native people lived in the region [1]. Salta city (SC) was founded in 1582 in the Lerma Valley, 1152 m above sea level. Including its metropolitan area, it has a population of 535,303 inhabitants [2]. After its foundation, the city grew exponentially, due to agriculture and mineral resources and its strategic position for trade between Lima and Buenos Aires until 1816 when the area was submerged in an economic crisis after the war of independence, a condition that lingered throughout much of the nineteenth century. However, in the early twentieth century, the arrival of Italian, Spanish, and Arab immigrants, together with the arrival of the railroad, revived trade and agriculture, and the economy of the zone improved [3]. The Calchaqui Valleys (CV), 1700–3000-m altitude in the Andes, cross the provinces of Catamarca, Tucumán, Jujuy, and Salta. The Diaguitas inhabited this area in the pre-Hispanic era. There is little historical information concerning these populations either before or after contact with the Europeans. Therefore, the exact origin and/or degree of admixture of present CV inhabitants is unknown, but a large proportion of the original population probably disappeared due to their particular history: invasion of the Incas, European colonization, and the policy of estrangement of the rebels, from the sixteenth to the late seventeenth centuries [4]. The current population (approximately 36,000 inhabitants) [2] has a low density and is unequally distributed, with Cafayate (approx. 15,000 inhabitants) and Cachi (7000) as the most populated localities (Supplementary Figure 1). Therefore, the present inhabitants of both Salta city and Calchaqui Valleys can be considered “mestizo” populations, the result of a mixture between immigrants of diverse origins, African slaves introduced in the colonial period, and natives of Amerindian tribes through a complex process of conquest and colonization of north-western Argentina (NWA).

The X-chromosome was chosen since it has proved to be an efficient tool in population genetics [5] and forensic practice [6], especially in complex and deficiency cases. Although population data on X-chromosome markers in Argentinian populations can be found in the literature [e.g., 7–10], this is a pioneer study on X-polymorphisms in the NWA region, focusing on the comprehensive analysis of 21 X-chromosomal markers of different types—Alu insertions and STRs—in two Salta province populations, aiming to (i) explore the variation of these polymorphic X-chromosome markers in the population of Salta city, as a reference of an urban population in NWA, and in the Calchaqui Valleys, as a rural population; (ii) correlate this genetic variation with the origin and migration history of these populations; and (iii) explore the usefulness of these markers in a forensic context.

Blood samples were obtained from 178 unrelated individuals (78 males and 100 females) from Salta province after informed consent: 73 living in different villages of the Calchaqui Valleys and 105 in Salta city. Samples were typed for two sets of X-chromosome genetic markers: (i) a set of nine X-chromosome Alu insertions (Ya5DP62, Yb8DP49, Yd3JX437, Yb8NBC634, Ya5DP77, Ya5NBC491, Yb8NBC578, Ya5DP4, and Ya5DP13) described by Callinan et al. [11] and (ii) 12 X-STRs included in the Investigator Argus X-12 kit (Qiagen GmbH, Hilden, Germany). Genotyping and statistical methods were performed as described in Ferragut et al. [12]. Proficiency testing of the Spanish and Portuguese Speakers Working Group of the International Society for Forensic Genetics (GHEP-ISFG, https://ghep-isfg.org) was conducted as a quality control. New variant alleles were sequenced by the Sanger method and aligned using the CodonCode Aligner program v.7.1.2 (CodonCode Corporation, Dedham, USA).

Allele frequency data in the studied populations (Salta city and Calchaqui Valleys) are included in Supplementary Tables 1–2. Exact test of population differentiation showed no significant differences between SC and CV. Values of average gene diversity across loci are summarized in Supplementary Tables 3–5. All analyzed loci were in the Hardy-Weinberg equilibrium after the Bonferroni correction for multiple tests (p > 0.00238) except for markers DXS10148 and Ya5DP62 in the Calchaqui Valleys population, due to a lack of heterozygotes. Five X-Alu insertions (Ya5DP62, Yb8DP49, Yd3JX437, Ya5DP77, and Ya5DP13) were revealed to be polymorphic in both populations, whereas the others appeared as monomorphic in at least one studied population: Ya5NBC491 and Yb8NBC578 for the insertion in both populations and Yb8NBC634 in CV; and for the absence of insertion, Ya5DP4 in the Calchaqui Valleys. Most polymorphic Alu elements showed moderate to low diversity. Ya5DP77 displayed the highest heterozygosity (0.461) and Ya5DP13 the lowest (0.020). These results agree with those found in other South American populations [8, 10, 13]. Average gene diversity for this set of X-chromosome Alu insertions was 0.157, an intermediate value between African and European populations and similar to Asian and Amerindian populations [11, 13]. All X-STRs were highly polymorphic in all populations (average gene diversity was 0.783 in VC and 0.798 in SC). Locus-by-locus analyses revealed DXS10135 had the greatest diversity (with 26 alleles and heterozygosity of 0.931), while DXS8378 and DXS10103 were the least diverse markers (with mean heterozygosity lower than 0.676). No identical haplotype-like allelic combinations of the 12 X-STRs markers were found when typing the 78 males from Salta province, although a higher number of samples would better support these results. Linkage groups (LG) 1–4 revealed 67, 60, 39, and 59 haplotypes, respectively (Supplementary Table 6). Of all the haplotypes observed, 76.9% were observed in only one individual, the most common one in 11 males in LG3, displaying a frequency of 0.141. In these populations, LG1 proved to be the most polymorphic group and LG3 the least variable. Notably, the rural population (CV) had slightly lower values in average gene diversity over loci and also in haplotype diversity (0.511 and 0.964, respectively) than the urban SC population (0.525 and 0.990, respectively). Pairwise LD analysis, after the Bonferroni correction for multiple tests, showed significant associations between Yb8NBC634-Yd3JX437 and Yb8NBC634-Ya5DP77 Alu pairs and DXS10103-HPRTB, DXS10103-DXS10101, and DXS10101-HPRTB pairs of STRs (Supplementary Table 7). STR loci associations involved markers located in the same linkage group (LG3); significant LD has also been reported for the same pairs of STRs in other studies [e.g., 12, 14].

During X-STR profiling, new (previously undescribed) alleles were detected in DXS10079, DXS10134, DXS7132, and DXS10148 markers. Traditional Sanger sequencing was performed to identify the structure of these new variants (Supplementary Table 8). Allele 13 in DXS10079 displayed nine straight repeats of the motif AGAA, instead of the 10–21 repeats in the reference structure described by Hering et al. [15]. In DXS10134, alleles 36.1, 37.1, 38.1, and 39.1 were detected, revealing the sequence variations responsible for these intermediate alleles were all due to one additional nucleotide (A) located outside the core repeat in the immediate upstream flanking region when compared to the sequence in Edelman et al. [16]. Allele 39.1 has also been described in a West African sample set [17], but the structure was different: in Guinea-Bissau samples, the additional base (A) was inserted in the last block of GAAA repeats (see Supplementary Table 8). Sequencing methods usually reveal that alleles derived from conventional PCR-CE systems have the same lengths but different sequences, leading to propose a new system of STR allele nomenclature for massively parallel sequencing of forensic STRs [18]. DXS7132 is a locus with a simple tandem of CTAT repeats [19], with intermediate alleles (16.3, 17.3, and 18.3) with a single-base deletion interrupting the perfect repeat motif: (CTAT)n-CAT-(CTAT)2. These intermediate alleles have been reported only in individuals with South American origin; this population specificity suggests a possible Amerindian origin of these alleles. In the present study, along with alleles 16.3 and 17.3, a previously undescribed allele with a fragment size corresponding to allele 13.3 was found. Sequencing data displayed the same deletion, but in a different repeat: (CTAT)7-CAT-(CTAT)6 (Supplementary Table 8). A highly complex compound structure was revealed for DXS10148. Sequencing data of Salta individuals displayed four different structures. The first, in allele 18 and the new variant 28, was according to Hundertmark et al. [20]. In intermediate alleles 18.1, 24.1, and 27.1 and in one individual with allele 26.1, a single-base insertion (A) interrupted the AAGG repeats: (GGAA)4-(AAGA)n-A-(AAGA)-(AAAG)4-N8-(AAGG)2-AAAG-(AAGG-AAAG)2-GGAAA. Sequencing of another individual with a fragment corresponding to allele 26.1 displayed a nucleotide between repeat motifs (AAGA)n and (AAAG)3, like individuals with alleles 22.1 and 28.1. Finally, the presence of an additional copy of the AAGG-AAAG motif was detected in two samples: one individual with the common allele 31 and one with a new variant, with a fragment size corresponding to allele 25. Regarding Alu insertion in locus Ya5NBC49, fragments with different lengths were detected (Supplementary Figure 2), with 73% of individuals having a band of approximately 415 bp rather than the expected 435 bp band [11]. Sequencing results revealed that length variation was due to a deletion of 20 bases in the poly (A) tail.

Statistical parameters of forensic interest were calculated for each polymorphic marker and population for each set of markers and across the 21-marker set (Supplementary Tables 3–5). As expected, considering the different nature of the polymorphisms, Alu insertion polymorphisms were considerably less informative than STRs. This 21 X-loci set proved to be suitable and highly informative for forensic casework in both populations studied, with values of combined mean exclusion chance (MEC) exceeding 0.999999 for duos and 0.99999999 for trios. In both populations, the combined power of discrimination (PD) yielded values greater than 1 in 4.504 E + 15 in females, and 1 in 1.829 E + 9 in males (Supplementary Table 5). The rural population (CV) had slightly lower values in all forensic efficiency parameters than the urban population (SC), in accordance with genetic diversity values.

To evaluate the genetic relationship between Salta province and other populations, pairwise FST genetic distances were computed (Supplementary Tables 9 and 10) and represented in an MDS plot (Supplementary Figure 3) separately for the two types of polymorphisms, since no other populations were studied with the full set of 21 X-chromosome markers. Population clustering reveals the ability of both sets of X-markers to clearly discriminate between continents, with European, Amerindian, and African populations distantly positioned. Using available data for these markers, North African groups are positioned much closer to the Europeans than to sub-Saharan populations. Regarding Amerindian populations, they show up on the left of the x-axis in the MDS plot for X-STRs but do not form a tight cluster, not even the ones from the same geographical region, suggesting differentiation by genetic drift in small and isolated populations [e.g., 10, 21]. For the X-Alu set of markers, only two Bolivian Amerindian populations were studied, showing close positioning. Salta and Calchaqui Valleys appear in the center of the plot, between Amerindian and European populations, for both sets of markers (Supplementary Figure 3), in accordance with the admixture from European and Native American ancestries of these populations. Using X-STR markers, a greater number of Amerindian populations were analyzed, allowing better definition. The Calchaqui Valleys were observed to stand closer to Amerindian populations while Salta city is closer to other Latin American populations.

Different admixture proportions from African, European, and Native American ancestries have contributed to the genetic heterogeneity found in Argentinean and Latin American populations in general. The highest Amerindian ancestry is found in areas with historically larger native populations, such as Andean regions and meso-America, where major pre-Columbian civilizations developed, whereas African ancestry is low in most of these populations [22, 23]. In Argentina, rural populations tend to show a greater predominance of the ancestral Native American substrate than urban ones, especially in maternal lineages [24,25,26], probably mirroring the strong sexual bias of the immigrants who settled during colonial times. Correspondingly, the X-polymorphism results in the present study show the CV population stands closer to the Amerindian cluster than SC, thereby suggesting rural populations from Salta province have maintained more important Native American ancestry in their genetic pool. Considering the positive correlation described between heterozygosity and European ancestry, the lower European component in CV, together with their rural and isolated origin, could explain the lower diversity values in CV versus the Salta urban population, as found in other Latin American populations [23].

The existence of a clear population stratification in Argentina [21, 27] emphasizes the need not to ignore this genetic heterogeneity in routine forensic casework and, hence, to develop local databases. This study provides a useful database of 21 X-chromosome markers, especially powerful for the X-STR set, for populations in Salta province (north-western Argentina). New detected variants are also described, confirming that some tandem repeat structures are more complex and variable than initially observed for this set of markers [28]. Results conclude both Salta populations harbor a high Native American ancestry component, as previously described for Argentina’s north-western region [27, 29], despite their self-recognized European ancestry.