Distribution of Killer-Cell Immunoglobulin-Like Receptor Genes and Combinations of Their Human Leucocyte Antigen Ligands in 11 Ethnic Populations in China

The aim of this study was to analyze the distribution of killer-cell immunoglobulin-like receptor (KIR) genes and their human leucocyte antigen (HLA) ligand combinations in different original ethnic populations in China, and thus, to provide relevant genomic diversity data for the future study of viral infections, autoimmune diseases, and reproductive fitness. A total of 1119 unrelated individuals from 11 ethnic populations—including Hani, Jinuo, Lisu, Nu, Bulang, Wa, Dai, Maonan, Zhuang, Tu, and Yugu—from four original groups, were included. The presence/absence of the 16 KIR loci were detected, and the KIR gene’s phenotype, genotype, and haplotype A and B frequencies, as well as KIR ligand’s HLA allotype and KIR–HLA pairs for each population, were calculated. Principal component analysis and phylogenetic trees were constructed to compare the characteristics of the KIR and KIR–HLA pair distributions of these 11 populations. In total, 92 KIR genotypes were identified, including six new genotypes. The KIR and its HLA ligands had a distributed diversity in 11 ethnic populations in China, and each group had its specific KIR and KIR–HLA pair profile. The difference among the KIR–HLA pairs between northern and southern groups, but not among the four original groups, may reflect strong pressure from previous or ongoing infectious diseases, which have a significant impact on KIR and its HLA combination repertoires.


Introduction
Segregated in different chromosomes, 6p21 and 19q13.4, human leucocyte antigen (HLA) and killer-cell immunoglobulin-like receptor (KIR) genes, respectively, exhibit diverse polymorphisms and their molecular expressions interact with each other as receptor ligands to ensure the proper role of nature killer (NK) cells in modulating an immune response [1,2]. Several studies on the coinheritance of these two genetic systems have indicated that carrying the appropriate KIR-HLA combination is important for human survival [3,4]. During human migration outward from Africa and the successive colonization worldwide, the cooperative KIR haplotypes and activating KIR-HLA pairs are important for humans to adapt to quickly changing environments and to increase population reproduction [5,6].

KIR Genotyping
The 16 KIR genes were genotyped using the Luminex MultiAnalyte Profiling System (xMAP) with a One Lambda KIR typing kit (One Lambda, Canoga Park, CA, USA), as previously reported [18]. Briefly, three separate PCR products were amplified: exon 3, exon 5, and exons 7-9. The PCR products were run on 2% agarose gel to confirm the specificity and efficiency of the reactions. Then, the PCR amplicons were denatured and hybridized with complementary 81-nucleotide oligonucleotide probes that had been immobilized on fluorescent-coated microsphere beads. At the same time, the biotinylated PCR products were labeled with phycoerythrin-conjugated streptavidin and immediately examined with the Luminex 200 system (Luminex, Austin, TX, USA). Genotype determination and data analysis were performed automatically using the LABScan 100 platform (One Lambda, Canoga Park, CA, USA) in accordance with the manufacturer's instructions. Genomic DNA was extracted from peripheral lymphocytes using a QIAamp Blood Kit (Qiagen, Hilden, Germany), in accordance with the manufacturer's protocol. DNA samples were quantified with a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, WI, USA) and adjusted to a concentration of 20 ng/µL.

KIR Genotyping
The 16 KIR genes were genotyped using the Luminex MultiAnalyte Profiling System (xMAP) with a One Lambda KIR typing kit (One Lambda, Canoga Park, CA, USA), as previously reported [18]. Briefly, three separate PCR products were amplified: exon 3, exon 5, and exons 7-9. The PCR products were run on 2% agarose gel to confirm the specificity and efficiency of the reactions. Then, the PCR amplicons were denatured and hybridized with complementary 81-nucleotide oligonucleotide probes that had been immobilized on fluorescent-coated microsphere beads. At the same time, the biotinylated PCR products were labeled with phycoerythrin-conjugated streptavidin and immediately examined with the Luminex 200 system (Luminex, Austin, TX, USA). Genotype determination and data analysis were performed automatically using the LABScan 100 platform (One Lambda, Canoga Park, CA, USA) in accordance with the manufacturer's instructions.

Statistical Analysis
Hardy-Weinberg's equilibrium for each of the alleles was assessed using the Guo and Thompson method [19]. For KIR genes, the observed frequency for each KIR gene was determined via direct counting and corresponded to the ratio of the number within the population that carried the gene to the total population number. KIR locus gene frequencies (KLFs) were estimated by using the formula KLF = 1 − 1 − f , where f is the observed frequency of a particular KIR sequence in a population. The genotypes were defined by referring to the Allele Frequencies website (http: //www.allelefrequencies.net). Each genotype was named in accordance with the genotype number. The genotypes that could not be found in the database were named as unknown. Group A and B haplotypes and frequencies were predicted on the basis of a previous study [20].
HLA genotyping of 11 ethnic populations has been reported previously [21][22][23][24][25], and the allelic frequencies are summarized in Supplementary Table S2. The frequencies of HLA-C1/C2 allotype, HLA-Bw4 (Bw4-80I and Bw4-80T)/Bw6 allotype, and HLA-A11, -A3 were calculated using direct counting. The HLA-C1 and HLA-C2, HLA-A11 and HLA-A3, and HLA-Bw4/Bw6 groups were used to analyze KIR-HLA combinations. The observed frequencies of KIR-HLA matched pairs were calculated using direct counting. The significance of the correlations of KIR and HLA frequencies among populations were estimated using the correlation coefficient (r) and the t-test was used to establish whether the correlation coefficient was significant using SPSS 16.0 [26]. The chord distance of Nei (Da distances among the populations were calculated based on 11 KIR gene (2DL1, 2DL2, 2DL3, 3DL1, 2DS1, 2DS2, 2DS3, 2DS4, 2DS5, 3DS1, and 2DL5) frequencies; HLA-A, HLA-B, and HLA-C allele frequencies; or HLA-KIR combinations. A neighbor-joining (NJ) tree was constructed using Mega 7.0 software based on the DA distance [27]. Principal component analysis (PCA) was also performed based either on KIR genes, HLA alleles, or KIR-HLA combination frequencies using SPSS 16.0 software [26]. Significant differences in KIR and KIR-HLA pair frequencies between two populations were determined using a contingency test. The difference between the northern and southern groups were detected using a t-test with SPSS 16.0 software [26]. A value of p < 0.05 was considered to be statistically significant. The observed 11 KIR gene (2DL1, 2DL2, 2DL3, 3DL1, 2DS1, 2DS2, 2DS3, 2DS4, 2DS5, 3DS1, and 2DL5) frequencies of 47 other populations were from previous studies and the DA distance among 58 populations were calculated (Supplementary Tables S5 and S6). The phylogenetic tree was constructed based on the DA distance using the minimum evolution method from Mega 7.0 software [28].

KIR Gene, Genotype, and Haplotype Frequencies
The observed KIR frequencies and the estimated gene frequencies for each locus in the 11 populations are listed in Table 1. The four framework loci (KIR3DL3, 3DP1, 2DL4, and 3DL2) were exhibited in all individuals in 10 populations, except one individual in Yugu, for whom 2DL4 and 3DP1 were not observed. The non-framework pseudogene 2DP1 was observed in all individuals in Hani, Nu, Dai, Zhuang, and Tu, but not in all other populations, with frequencies of 94.8%-99.1%. The frequencies of 3DL1, 2DL1, and 2DS4 were about 90%-100% in 11 populations, with the exception in Jinuo and Bulang, which showed frequencies of 88% and 81% at 3DL1 and 88% and 79% at 2DS4, respectively. Other activating KIRs, including 3DS1, 2DS1, 2DS3, and 2DS5, as well as inhibitory KIRs, including 2DL2 and 2DL5, exhibited diverse distributions in different populations. Bulang was different compared with the other 10 populations at 3DS1 and 2DS1 (p < 0.05), 9 other populations except Zhuang at 2DS3, and 9 other populations except Jinuo at 2DS4. The following difference was among Nu and others. For all populations, the most diverse was at 2DS3 (Supplementary Table S3).
In total, 92 KIR genotypes were identified, including 6 new genotypes: 3 in Tu, 2 in Jinuo, and 1 in Wa (Table 2). Genotypes 1, 2, 4, and 8 were observed in all the populations but showed diverse frequencies. Genotype 1 was predominant in all populations except in Bulang. On the contrary, the predominant genotypes in Bulang were genotype 8 followed by genotypes 1, 2, and 75, as previously reported [18]. Originating from the same Baipu ancient group, Wa did not show a distribution similar to Bulang, with genotype 1 being the most predominant, followed by genotypes 2 and 4. The frequencies of genotype 1 were as high as 0.679 in Nu, 0.510 in Hani, 0.469 in Yugu, 0.465 in Lisu, 0.432 in Zhuang, and 0.425 in Wa.   The frequencies of the group A and B haplotypes in the 11 populations were deduced from the genotype data ( Figure 2). As with most populations worldwide, the A haplotype was predominant. The frequencies of the A haplotype were around 0.657-0.830 in Nu, Hani, Wa, Lisu, Zhuang, Maonan, and Yugu, while they were around 0.576-0.593 in Tu, Jinuo, and Dai. In Bulang, the frequencies of the A and B haplotypes were almost equal (0.491 vs. 0.509). Haplotype differences were identified among Nu and 10 other populations except Jinuo; among Bulang and Hai, Lisu, Nu, Wa, Maonan, Zhuang, and Yugu; and between Hani and Jinuo, and Dai and Tu (Supplementary Table S3). The distributions of KIR genes, genotypes, and haplotypes did not show any consistency among their original ancient group or linguistic subfamily. The frequencies of the group A and B haplotypes in the 11 populations were deduced from the genotype data ( Figure 2). As with most populations worldwide, the A haplotype was predominant. The frequencies of the A haplotype were around 0.657-0.830 in Nu, Hani, Wa, Lisu, Zhuang, Maonan, and Yugu, while they were around 0.576-0.593 in Tu, Jinuo, and Dai. In Bulang, the frequencies of the A and B haplotypes were almost equal (0.491 vs. 0.509). Haplotype differences were identified among Nu and 10 other populations except Jinuo; among Bulang and Hai, Lisu, Nu, Wa, Maonan, Zhuang, and Yugu; and between Hani and Jinuo, and Dai and Tu (Supplementary Table S3). The distributions of KIR genes, genotypes, and haplotypes did not show any consistency among their original ancient group or linguistic subfamily.

HLA Allotype Frequencies
The frequencies of HLA-A11/A3, HLA-Bw4 (Bw4-80I and Bw4-80T), and HLA-C1 and HLA-C2 were calculated from the HLA-A, HLA-B, and HLA-C allele genotyping results (Table 3). HLA-A11/A3 were predominant in all the populations living in southern China, with frequencies higher than 0.556, and they accounted for around 80% of HLA-A alleles in Bulang, Wa, and Hani. On the contrary, HLA-A11/A3 were around 40% in Tu and Yugu living in northern China. HLA-Bw4 existed commonly in Tu and Yugu, with frequencies of 0.700 and 0.688, but only with frequencies of 0.283 in Bulang. HLA-C1 was observed in all individuals in Jinuo, Lisu, Maonan, and Zhuang, with frequencies around >95% in other southern Chinese populations, but with frequencies of 0.914 and 0.833 in Tu and Yugu, respectively, from northern China. On the contrary, the frequencies of HLA-C2 were around 50% in Tu and Yugu but was only 0.073 in Maonan. This HLA characteristic reflected the northern and southern Chinese original difference, which has been confirmed in previous studies. Therefore, we divided the present study populations into two groups: the southern group, which included Hani, Jinuo, Lisu, Nu, Bulang, Wa, Dai, Maonan, and Zhuang, and the northern group, which included Tu and Yugu. Differences between the northern and southern groups were observed (data not shown).

HLA Allotype Frequencies
The frequencies of HLA-A11/A3, HLA-Bw4 (Bw4-80I and Bw4-80T), and HLA-C1 and HLA-C2 were calculated from the HLA-A, HLA-B, and HLA-C allele genotyping results (Table 3). HLA-A11/A3 were predominant in all the populations living in southern China, with frequencies higher than 0.556, and they accounted for around 80% of HLA-A alleles in Bulang, Wa, and Hani. On the contrary, HLA-A11/A3 were around 40% in Tu and Yugu living in northern China. HLA-Bw4 existed commonly in Tu and Yugu, with frequencies of 0.700 and 0.688, but only with frequencies of 0.283 in Bulang. HLA-C1 was observed in all individuals in Jinuo, Lisu, Maonan, and Zhuang, with frequencies around >95% in other southern Chinese populations, but with frequencies of 0.914 and 0.833 in Tu and Yugu, respectively, from northern China. On the contrary, the frequencies of HLA-C2 were around 50% in Tu and Yugu but was only 0.073 in Maonan. This HLA characteristic reflected the northern and southern Chinese original difference, which has been confirmed in previous studies. Therefore, we divided the present study populations into two groups: the southern group, which included Hani, Jinuo, Lisu, Nu, Bulang, Wa, Dai, Maonan, and Zhuang, and the northern group, which included Tu and Yugu. Differences between the northern and southern groups were observed (data not shown).

KIR-HLA Combination
The frequencies of KIR3DL2 and HLA-A11/A3 were calculated first. Since the interaction of KIR2DS4 and KIR2DS2 with HLA-A11 has been demonstrated, their combinations have also been calculated [14,29]. The frequencies of KIR3DL2+A11/A3 and KIR2DS4+A11/A3 were from 0.417 to 0.879, while the frequencies of KIR2DS2+A11 were lower, with frequencies from 0.038 to 0.299 (Table 4). In Nu, the frequency of KIR2DS2+A11 was only 0.038, though HLA-A*11:01 was predominant, with a frequency of 0.411 (Table 4). The individuals carrying either KIR3DL1 or KIR3DS1 and its HLA-Bw4 ligands, or carrying both KIR3DL1 and KIR3DS1 together with its HLA-Bw4 ligands, were counted for the KIR-HLA combination ( Table 5). KIR3DL1/3DS1+Bw4 was commonly exhibited in all populations, with frequencies of 0.383-0.700, except in Bulang. The total frequency of the KIR3DL1/3DS1+Bw4 combination was 0.283, and the frequency of either KIR3DL1+Bw4 or KIR3DS1+Bw4 pairs was 0.057. The frequencies of 3DL1+3DS1+Bw4 were predominant in Bulang, and 3DL1+3DS1+Bw4 and 3DL1+Bw4 were almost similar in Jinuo; however, KIR3DL1+Bw4 was predominant in other 9 populations. In Hani and Wa, all individuals carried either 3DL1+Bw4 or 3DL1+3DS1+Bw4 together, and no one carrying only 3DS1+Bw4 was observed. Further analysis of KIR with the presence of isoleucine at position 80(Bw4-80I) as well as with the presence of threonine at position 80(Bw4-80T) was performed. One individual in Dai and Jinuo with HLA-Bw4 were unable to have their KIR ligands identified, two individuals in Lisu were not able to have their KIR ligands identified, while all the other individuals with KIR-Bw4 ligands were identified. For the 3DL1/3DS1+Bw4 pair, the frequencies of 80I3DL1+Bw4 80I were higher than 3DS1+Bw4 80I. Moreover, in Hani, Nu, Wa, and Maonan, no individuals who only carried 3DS1+Bw4 80I were observed. There were more individuals only carrying KIR3DL1+Bw4-80I than those only carrying KIR3DL1+Bw4-80T in Hani, Lisu, Bulang, and Yugu, while there were fewer in Jinuo, Dai, and Maonan, and there was no difference in Nu, Wa, Zhuang, and Tu. There was a similar finding for the KIR3DL1+Bw4-80T and KIR3DS1+Bw4-80T pairs. Further analysis of the southern and northern groups indicated that the frequencies of KIR3DL1+Bw4 were lower in the southern group than in the northern group (0.451 ± 0.120 vs. 0.659 ± 0.018, p = 0.001).

Phylogenetic Analysis
Both principal component analysis (PCA) and phylogenetic trees using KIR, HLA, and KIR-HLA combination frequencies were employed. For PCA based on 11 KIR (3DL1, 2DL1, 2DL3, 2DS4, 2DL2, 2DL5, 3DS1, 2DS1, 2DS2, 2DS3, and 2DS5) plots, Bulang showed distance from the other 10 populations (Figure 4a), and the other 10 populations did not cluster with their linguistic family as based on HLA-A, -B, and -C (Figure 4b), which agreed with previous studies. On the PCA plots based on KIR-HLA pairs, Yugu and Nu from northern China clustered together and showed distance from the other nine populations from southern China (Figure 4c).  On the phylogenetic tree based on KIRs, for the trees constructed either by all 11 KIR genes (Figure 5a), or by the inhibitor KIR genes or by the activating KIR genes (data not shown) had no clear clustering among the populations. For the NJ tree constructed using HLA genes, Tu and Yugu On the phylogenetic tree based on KIRs, for the trees constructed either by all 11 KIR genes (Figure 5a), or by the inhibitor KIR genes or by the activating KIR genes (data not shown) had no clear clustering among the populations. For the NJ tree constructed using HLA genes, Tu and Yugu clustered together as one major branch. Bulang and Wa of Khmer clustered together, and Maonao, Zhuang, and Dai of Daic clustered together with Jinuo (Figure 5b). When 58 populations were compared on the phylogenetic tree using 11 KIRs frequencies, most populations clustered together according to their geographic location of Asian, European, African, and American ( Figure 6). However, the closeness were not displayed clearly as in the NJ tree constructed by the HLA genes, in which the populations clustered according to their evolutional relationship [25]. In the Asian branch, on one hand, most Han populations in northern China clustered with Japanese and Korean, but also together with Nu, Yi, and   On the phylogenetic tree based on KIRs, for the trees constructed either by all 11 KIR genes (Figure 5a), or by the inhibitor KIR genes or by the activating KIR genes (data not shown) had no clear clustering among the populations. For the NJ tree constructed using HLA genes, Tu and Yugu clustered together as one major branch. Bulang and Wa of Khmer clustered together, and Maonao, Zhuang, and Dai of Daic clustered together with Jinuo ( Figure 5b). When 58 populations were compared on the phylogenetic tree using 11 KIRs frequencies, most populations clustered together according to their geographic location of Asian, European, African, and American ( Figure 6). However, the closeness were not displayed clearly as in the NJ tree constructed by the HLA genes, in which the populations clustered according to their evolutional relationship [25]. In the Asian branch, on one hand, most Han populations in northern China clustered with Japanese and Korean, but also together with Nu, Yi, and     Figure 6. Neighbor-joining tree constructed using 11 KIR genes frequencies of 58 populations worldwide. The optimal tree was one with the sum of branch length = 0.159.

Discussion
The extensive diversity of HLA and KIR genes and their interactive roles in the immune response make these genes coevolve in genotypic combination. Disease and population studies have confirmed that they evolved together as specific KIR-HLA pairs to regulate NK cell function and play a vital role in the innate defense against pathogens and early placentation [30][31][32]. KIR-HLA combination studies have been performed in different populations worldwide; however, previous studies in China were limited to the Han population [4,[33][34][35][36][37][38]. In the present study, we analyzed KIR and its HLA pairs in 11 ethnic populations from northern and southern China covering four different linguistic language families that represent the major origins of Chinese ethnic populations. This study not only provides useful genomic diversity data for the future study of viral infections, autoimmune diseases, and reproductive fitness among these populations, but also reveals clues about HLA and KIR interaction and coevolution under the diverse change of pathogen infections.
The coevolution of HLA and KIR has been proved by several population studies worldwide, and different KIR-HLA pair correlations were identified in different populations. In 2006, Single et al. studied the distribution of KIR and its HLA ligands in 30 populations worldwide and observed that a balancing selection acted on the negative correlation between KIR3DS1 and HLA-Bw4-80I pairs [30]. In 2013, Hollenbach et al. compared KIR-HLA pairs in 105 populations worldwide and revealed a significant correlation between KIR2DL3 and HLA-C ligands. However, the correlation for KIR3DL1 and HLA-Bw4 pairs was not significant [39]. In the Italian population, a correlation between KIR and HLA-C ligands was not observed; instead, a correlation between KIR3DL1 and HLA-Bw4 ligands, as well as KIR3DL2 and HLA-A3 and HLA-A11 ligands, was observed [40]. In the present study, the observed frequencies for KIR2DL3 and HLA-C1 ligands showed a correlation (r = 0.637, p = 0.035) that agreed with Hollenbach et al.'s study, as well as Gendzekhadze's study in the Yucpa og South Amerindian [4]. The association of KIR2DL3 and HLA-C1 has also been investigated regarding the Hepatitis C virus and Malaria infection [41,42]. In Hirayasu et al.'s study, they found KIR2DL3+HLA-C1, but no other KIR-HLA pairs were associated with cerebral malaria, and the frequency of combination was significant lower in malaria high-endemic populations. This result suggested that natural selection has reduced the KIR2DL3+HLA-C1 frequencies in malaria high-endemic populations to favor the development of malaria [41]. Thus, KIR-HLA coevolution may be driven by microbial pathogens, resulting in specific distributions of KIR-HLA pairs in different populations.
The genetic difference between southern and northern Chinese has been confirmed in studies of HLA [35,43], as well as immunoglobulins [44], microsatellites [45], and Y-chromosome single-nucleotide polymorphisms [46]. Furthermore, population migration from northern to southern China has frequently happened throughout Chinese history [15][16][17]. In the present study, KIR2DL2/3+HLA-C1, KIR2DL1+HLA-C2, and KIR2DS1+HLA-C2 pairs showed clear differences between the northern and southern groups. The frequencies of KIR2DL2/3+HLA-C1 pairs were significantly higher in the southern group than in the northern group (0.978 vs. 0.860, p = 0.00005). This difference has also been investigated in Han population. The frequencies of KIR2DL2/3+HLA-C1 pairs were higher in two southern Chinese Han, namely Guangdong Han and Yunnan Han, than in the northern ethnic group, with frequencies of 0.981 and 0.950, respectively. In contrast, the frequencies of KIR2DL1+HLA-C2 and KIR2DS1+HLA-C2 pairs were lower than in the northern group, with frequencies of 0.294 and 0.351 for KIR2DL1+HLA-C2, and 0.100 and 0.155 for KIR2DS1+HLA-C2 in Guangdong Han and Yunnan Han, respectively [36,38]. Furthermore, the Tu and Yugu formed a cluster and showed a distance from other ethnic populations in southern China in the PCA plot constructed using KIR-HLA pair frequencies.
According to historical records, northern and southern China underwent different pathogenic pressures. Regarding malaria (a serious infectious disease prevalent in China since 2700 BC), its epidemic area was focused in southern China, whereas northern China was malaria free or had a very low incidence rate [47,48]. Historically, Yunnan Province has been the most high-risk malaria area, especially along the China-Myanmar border [49,50]. In the present study, Tu and Yugu are from northern China, while the other populations are all in southern China, the most high-risk malaria areas. Therefore, we deduced that the different distribution between the northern and southern groups in China may have been caused by severe infectious disease epidemics, such as malaria.
Except for KIR+HLA-C pairs, KIR+HLA-Bw and KIR+HLA-A3/A11 pair differences were diverse in different populations. When considering KIR genes, genotypes, and haplotypes, more diversity was exhibited. It is interesting to note that the HLA gene distributions were in accordance with the population linguistic group and their origins. Both in the PCA plot and neighbor-joining tree constructed using HLA allele frequencies, the same linguistic origin populations clustered together. However, there were no clear cluster trends to distinguish the populations according to their origin or linguistic classification using either KIR, activated KIR, or KIR genotype frequencies among 11 ethnic populations. Moreover, in the phylogenetic tree constructed using 11 KIRs frequencies of 58 populations worldwide, there was a geographic closeness among Asians, Africans, Europeans, and Americans, while the 11 ethnic populations in the present study still did not show a clear origin or linguistic clustering trend. Compared with HLA genes, KIR genes have experienced a rapid evolution through a combination of gene duplication and nonhomologous recombination [1]. The extensive diversity of KIR genes in different populations worldwide indicates that distinct diseases have recently acted or are still acting to select on KIR repertoires [2]. This evolution was thought to be driven by the selective pressure of pathogen invasion, as well as reproduction. Moreover, haplotypes A and B are thought to have maintained a balance selection in human beings. The A haplotypes are associated with an improved response to pathogens, while B haplotypes are associated with reproductive fitness [51,52]. Previous studies have indicated that the populations are related to their geographic distribution based on KIR haplotype B but do not show a correlation based on haplotype A [53]. Moreover, it has been reported that B haplotypes are more prevalent in Australian Aborigines and Asian Indians, where the possible reason is due to these populations maybe being under strong pressure from infectious diseases [2]. In the present study, the frequencies of haplotypes A and B were almost similar to each other in Bulang, which showed a difference from other Asian populations. On the contrary, in Nu, haplotype A was as high as 0.830, which showed a significant difference from the other 10 populations. This extensive range of haplotype A in the present study, from 0.491 in Bulang to 0.830 in Nu, together with the diverse frequencies worldwide, may be the result of a founder effect, genetic drift, or natural selection [5,31]. Therefore, the distribution of KIR profiles among the present study populations could not be interpreted as a phylogenetic tree.
In conclusion, the distribution of KIR and its HLA ligands in 11 ethnic populations in China exhibited diverse characteristics, where each group had its specific KIR and KIR-HLA pair profile. The difference of KIR-HLA pairs between the northern and southern groups, but not among the four original groups, may reflect the strong pressure from previous or ongoing infectious diseases that have had a significant impact on KIR and its HLA combination repertoires.