Forensic efficiency estimate and phylogenetic analysis for Chinese Kyrgyz ethnic group revealed by a panel of 21 short tandem repeats

2.1. Sample collection and DNA extraction

Peripheral blood was extracted from 307 unrelated healthy individuals dwelling in the Kizilsu Kirghiz Autonomous Prefecture of Xinjiang Uygur Autonomous Region for more than three generations. Written informed consent was obtained from every participant. After collection of the peripheral blood, a small part of the blood, which was spread on a fresh filter and allowed to dry at room temperature, was made into a bloodstain for long-term conservation, and the remaining part was frozen for storage. This study was carried out according to the humane and ethical research principles approved by the ethical committee of Xi'an Jiaotong University Health Science Center, China (no. XJTULAC201). DNA was extracted from the bloodstain mentioned above by the Chelex-100 method [15].

2.2. Polymerase chain reaction amplification and short tandem repeat typing

The 21 non-CODIS STR loci were amplified simultaneously in a single polymerase chain reaction (PCR) system by the AGCU 21+1 fluorescence amplification reagents (AGCU ScienTech Incorporation, Wuxi, Jiangsu, China) with a GeneAmp PCR System 9700 Thermal Cycler (Applied Biosystems, Foster City, CA, USA). The PCR conditions could be obtained from the developmental validation of the reagent [9]. STR genotyping was performed on an ABI PRISM 3130 Genetic Analyzer (Applied Biosystems) and analysed by the Genemapper ID 3.2 software (Applied Biosystems).

2.3. Quality control

The study was in accordance with ISFG recommendations by Schneider [16] on the analysis of DNA polymorphisms.

2.4. Statistical analyses

A modified Powerstat v. 1.2 spreadsheet [17] was used to compute the allelic frequencies and forensic parameters including matching probability (MP), power of discrimination (PD), probability of exclusion (PE), observed heterozygosity (HO), polymorphism information content (PIC), typical paternity index (TPI) and exact tests of the Hardy–Weinberg Equilibrium (HWE). Meanwhile, the values of expected heterozygosity (HE), cumulative power of discrimination (CPD) and cumulative probability of exclusion (CPE) were calculated with formulas directly. Linkage disequilibrium (LD) exact tests of each pair of STR loci and pairwise F_st values were estimated by Genepop v. 4.0.10 (http://genepop.curtin.edu.au/). Population structure analyses were conducted among the studied Kyrgyz ethnic group and other populations previously published using the Structure program v. 2.2 (http://pritch.bsd.uchicago.edu/structure.html), and the plot of optimum K (K = 6) determined by Structure Harvester v. 0.6.94 [18] was portrayed using Distruct v. 1.1 (https://web.stanford.edu/group/rosenberglab/distruct.html). The neighbour-joining (NJ) tree was constructed based on the D_A values computed with the Dispan program by MEGA v. 6.06 (http://megasoftware.net/) and another tree was constructed by Phylip v. 3.69 based on allelic frequencies. PASW Statistics v. 18 (http://www.winwrap.com) was used to create a multidimensional scaling (MDS) plot based on pairwise F_st values. In addition, locus-by-locus F_st and p-values between the Kyrgyz group and the other compared populations were calculated by Arlequin software v. 3.1 (http://cmpg.unibe.ch/software/arlequin3).

3.1. Genetic polymorphism analyses of 21 non-combined DNA index system loci

Allelic frequencies and forensic parameters of the 21 non-CODIS STR loci are listed in tables 1 and 2, respectively. There were 168 alleles observed with corresponding allelic frequencies from 0.0016 to 0.4788 in the group. No significant deviations (p > 0.05) were observed from the HWE in these 21 STR loci. The maximum values of PD, PE, PIC and TPI were 0.9494, 0.7081, 0.8090 and 3.4886, respectively, which were all observed at locus D19S433. On the contrary, the minimum values all observed at locus D1S1627 were 0.7617, 0.2565, 0.51044 and 1.1629, respectively. In addition, the MP values ranged from 0.0506 at locus D19S433 to 0.2383 at locus D1S1627. HO values were in a range from 0.8567 (D19S433) to 0.5700 (D1S1627), while the HE values ranged from 0.8284 (D19S433) to 0.5955 (D1S1627). The CPD and CPE values of all these 21 STR loci were 0.99999999999999999998844 and 0.9999993992, respectively. The results mentioned above indicated that the panel of non-CODIS 21 STR loci could be used as a sensitive and accurate tool in routine forensic caseworks of the studied Kyrgyz group.

3.2. Linkage disequilibrium analyses of pairwise short tandem repeat loci

A previous study has verified that the physical distances among the non-CODIS 21 STRs were all greater than 10 Mb, that is, these loci were not linked with each other [9]. Also in this research, LD analyses should be conducted between each pair of STR loci before the next step in the analyses of population genetics and forensic science. Therefore, the LD tests in the 210 pairs of all the STR loci were performed by Genepop v. 4.0.10. After Bonferroni correction (p = 0.05/210 = 0.00024), no LD was observed in all the pairwise comparisons. Thus, the 21 STR loci could be considered as independent loci in the following analyses.

3.3. Structure analyses based on raw population data

The population structure analyses were conducted on the basis of the raw population data of the same 21 STR loci between the studied Kyrgyz and other previously published populations by the Structre program. All the reference populations included nine populations from East Asia (Ningxia Han [19], Guanzhong Han [20], Tibetan [21], Bai [22], Yi [23], Russian [24], Salar [25], Tujia [26] and Mogolian [27]) and three populations from Central Asia (Chinese Kazak [28], Uygur [29] and the studied Kyrgyz). Each of K = 2–7 with 15 runs was carried out; then the optimum K was selected by Structure Harvester v. 0.6.94 and the results are shown in the electronic supplementary material, figure S1, suggesting that K = 6 was the most appropriate configuration. As shown in figure 1, all the analysed populations had a similar pattern of components distribution at K = 6, and proportions of the six colours were almost equal. Even the three populations (Kazak, Uygur and the studied Kyrgyz ethnic groups) from Central Asia were consistent with the nine East Asian populations roughly. Therefore, by the panel of 21 non-CODIS STR loci, no obvious population structure differentiation was observed among the Kyrgyz and the 11 reference populations mentioned above.

3.4. Interpopulation differentiations

The locus-by-locus F_st and p-values were calculated by the analysis of molecular variance (AMOVA) method between the Kyrgyz group and the 11 reference populations using Arlequin software v. 3.1. The results are given in table 3. Significant differences (p < 0.05) were detected between the Kyrgyz and the Yi group at 14 loci, followed by the Salar, Tibetan, Ningxia Han and Guanzhong Han groups at 11 loci, the Bai group at nine loci, the Tujia group at eight loci, the Russian and Mogolian groups at seven loci, the Uygur group at six loci, and finally the Kazak group at only one locus. Highest ethnic diversity was detected at the three loci (D10S1248, D12ATA63 and D1S1627) with significant differences observed between the Kyrgyz and the other 10 groups; in contrast, the lowest ethnic diversity was detected at locus D6S474 with no significant differences observed between the Kyrgyz group and the other reference groups.

3.5. Genetic distances and population differentiations

For further study, the pairwise D_A and F_st values between the Kyrgyz and the other reference groups were calculated, which are not only presented in table 4 but also shown with a clustered bar chart (electronic supplementary material, figure S2). The largest two values of D_A were observed between the Kyrgyz and the Yi group (0.0356) and then the Russian group (0.0252), whereas the smallest two were found between the Kyrgyz and the Uygur group (0.0083), and then the Kazak group (0.0097). Correspondingly, the F_st values were ranged from 0.0002 (between the Kyrgyz and the Kazak group) to 0.0350 (between the Kyrgyz and the Russian group), which were basically in line with the D_A values. The parameters in table 4 directly demonstrated that the studied Kyrgyz ethnic group and the two Central Asian populations (Kazak and Uygur ethnic groups) had close genetic relationships, which contrasted with the Yi and Russian groups, with relatively distant genetic relationships.

3.6. Phylogenetic analysis of 12 populations

The NJ tree was constructed by MEGA v. 6.06 based on D_A values represented in figure 2a. From this figure, the Kyrgyz group labelled by a green dot was first clustered with the Uygur and the Kazak group (green dots), and then with six East Asian ethnic groups marked by red dots: Tibetan, Tujia, Bai, Ningxia Han, Guanzhong Han and Mogolian group; however, there was another cluster that consisted of three East Asian ethnic groups marked by red dots: Yi, Russian and Salar groups. In figure 2b, another NJ tree was constructed by Phylip v. 3.69 based on gene frequencies, which is consistent with figure 2a. In the case of using different software based on different data formats, the three groups from Central Asia were clustered together in the two NJ trees, which revealed that they had stronger genetic relationships than the other populations from East Asia.

3.7. Multidimensional scaling based on the pairwise F_st values

As shown in figure 3, MDS was performed among the 12 populations based on the pairwise F_st values and the studied Kyrgyz ethnic group was marked with a red colour. The result indicated that the 12 populations could be divided into three parts: Tibetan, Mogolian, Bai, Tujia, Guanzhong Han and Ningxia Han ethnic groups were clustered in the upper quadrant; Salar, Russian and Yi ethnic groups in the lower right quadrant; whereas, Kazak, Uygur and the studied Kyrgyz ethnic groups in the lower left quadrant. Compared with the nine East Asian ethnic groups, the Kyrgyz ethnic group had even more intimate relationships with the Kazak and Uygur ethnic groups, which indicated that the Kyrgyz group probably had close genetic relationships with the two ethnic groups from Central Asia.

According to historical records, from the Western Han Dynasty to the middle of the Qing Dynasty, the Kyrgyz group, mainly stemming from the Yenisai River to the Tianshan Mountains and Central Asia, experienced five westward migrations which were basically facilitated by warfare [30]. In this study, the Kyrgyz group residing in the southwestern part of the Xinjiang Uygur Autonomous Region, China, broadly assimilated the culture of the western regions after long-term dwelling with the Uygurs, Kazaks and other ethnic groups in Xinjiang.