Determination of major histocompatibility class I and class II genetic composition of the Caribbean Primate Center specific pathogen-free rhesus macaque (Macaca mulatta) colony based on massively parallel sequencing

Abstract

Background

Knowledge of MHC composition and distribution in rhesus macaque colonies is critical for management strategies that maximize the utility of this model for biomedical research.

Methods

Variation within the Mamu-A and Mamu-B (class I) and DRB, DQA/B and DPA/B (class II) regions of 379 animals from the Caribbean Primate Research Center’s specific pathogen free (SPF) colony was examined using massively parallel sequencing.

Results

Analyses of the seven MHC loci revealed a background of Indian origin with high levels of variation despite past genetic bottlenecks. All loci exhibited mutual linkage disequilibria while conforming to Hardy–Weinberg expectations suggesting the achievement of mutation-selection balance.

Conclusion

The CPRC’s SPF colony is a significant resource for research on AIDS and other infectious agents. Characterizing colony-wide MHC variability facilitates the breeding and selection of animals bearing desired haplotypes and increases the investigator’s ability to understand the immune responses mounted by these animals.

Introduction

As genes within the major histocompatibility complex (MHC) class I and class II regions facilitate prediction of the immune response, the use of MHC-defined macaques as research subjects improves the reproducibility and validity of experimental results and is thus a vital component of many NIH funded biomedical research programs. Figure 1 illustrates the various human MHC class I and class II regions including the genes coding for the human leukocyte antigens (HLA)-A, -B, -E, and -F orthologs exhibited by macaques [1–3]. As Figure 1 depicts, the macaque MHC class I genomic region has many more class I genes than the HLA-class I genomic region as a result of gene duplication events. For instance, rhesus macaques exhibit 4 fold more potentially functional class I proteins than humans [4]. While the macaque MHC class I region is far more complicated in its organization than the HLA-class I region of humans, the class II genes are more conserved between macaques and humans [5] (although many linked macaque MHC class I and II genes have yet to be adequately described [6]). The advent of massively parallel sequencing (MPS) technologies has aided the comprehensive examination and characterization of MHC organization of colony bred macaques [7]. However, there is no efficient protocol for breeding colony managers to compile this information and utilize it to maximize the long-term efficiency of MHC-defined colony production.

Figure 1.

Map of the macaque MHC regions (above), based on LDr estimates generated by the present study, compared to the orthologous human MHC region, also called the HLA (human leukocyte antigen) complex (below) adapted from [49] and [5]. An A001 haplotype is the summary of all A alleles as shown in Supplementary table 1. The same is true for the B and DRB haplotypes (Supplementary table I).

As part of its MHC screening program to provide larger cohorts of animals with enhanced MHC characterization to investigators, the Caribbean Primate Research Center (CPRC) has been testing its specific pathogen-free (SPF) rhesus macaques (Macaca mulatta) for the Mamu-A001, A002, A011, B001, B003, B008 and B017 haplotypes by genotyping MHC-linked short tandem repeat (STR) loci and/or performing sequence specific primer-polymerase chain reaction (SSP-PCR) amplification [8, 9]. Since some MHC haplotypes, groups of alleles of MHC genes that are co-inherited, serve a critical simian immunodeficiency virus (SIV) epitope presentation function in macaques that is associated with an adaptive cellular immune response in animals infected with SIV [10–12], individuals exhibiting these haplotypes are considered highly desirable in AIDS research. For instance, approximately 20% of Mamu-B*17-positive and 50% of Mamu-B*008-positive rhesus macaques can control replication of SIVmac239 spontaneously [13, 14]. Individuals with the Mamu-A*01 allele, which, like the Mamu B*01 allele, occurs in 20–30% of Indian rhesus macaques [15], can control viral replication better and survive longer than those without the allele [16, 17], and those with the class I alleles Mamu-A*01, Mamu-B*17, and Mamu-B*29 can control the replication of a highly pathogenic SIV clone [18]. Several class II haplotypes, including Mamu-DRB1 0306 and Mamu-DRB11003, have also been shown to be over-represented in virus-infected macaque elite controllers [21], and ongoing efforts in the search for effective therapeutics and vaccines against HIV/AIDS have also identified several CD8+ T epitopes that play a crucial role in responses against SIV [18, 19, 20]. As more haplotypes that influence the acute-phase cytotoxic T lymphocyte responses (CTR) and long-term clinical outcomes of SIV/HIV infection are identified [7], requests by investigators for information on these additional MHC haplotypes will also increase.

The CPRC is now relying on high throughput MHC sequencing to increase efficiency in characterizing additional class I and class II haplotypes. The introduction of macaque MHC genotyping via deep sequencing techniques [22, 23] has enhanced sample throughput while lowering per sample costs, which allows increased efficiency in MHC typing by generating complete haplotype libraries from significant numbers of colony animals. This study reports MHC data from the CPRC SPF colony that was recently generated using the deep sequencing method first described in Wiseman et al. [23]. The composition and distribution of MHC class I and class II haplotypes from the CPRC SPF colony described here will directly benefit the development of the center’s SPF colony for AIDS-related research as well as initiatives on infectious or autoimmune diseases in which the immune response of the MHC class I or II haplotypes play a crucial role.

Methods and Materials

MHC genotyping of a study cohort of 379 CPRC SPF rhesus macaques was performed by the Genetics Services Unit at the Wisconsin National Primate Research Center (WNPRC) on a fee for service basis. Genomic DNA (gDNA) extracted from frozen blood samples was amplified with a Fluidigm Access Array using a multiplex panel of PCR primers flanking the hypervariable (i.e., most polymorphic [7]) peptide binding domains encoded by exon 2 of class I (Mamu-A and -B) and class II (Mamu-DRB, -DQA, -DQB, -DPA and -DPB) loci (Figure 1). Over 20 million sequence reads were generated with an average of 53,000 reads per animal. The amplicons were cleaned-up, pooled and then sequenced on an Illumina MiSeq instrument [R. Wiseman, pers. comm. August 2017; September 2017] [24]. The Genetics Services Unit’s custom database of rhesus macaque sequences of major class I and class II alleles was used to manually map the sequence reads, each of which was tagged with a unique barcode that referenced the animal from which it was derived, to these reference sequences and assign them to the allele they most likely represent [R. Wiseman, pers. comm. August 2017; September 2017] [24]. These ‘major’ alleles that are defined by known haplotypes represent the vast majority of sequence reads and are assumed to initiate most of the immune responses [24].

Genotyping based on MiSeq sequence reads is limited because the sequence reads do not span entire open reading frames to allow an exact identification of alleles but rather rely on low-resolution definitions of lineages as families of closely related sequences that differ by only one or a few nucleotide differences [R. Wiseman, pers. comm. November 2017]. A consequence of the short amplicons is that the class II sequence data that have only been defined by MiSeq sequencing, which is the case for some DPA/DPB and DQA/DQB analyses, cannot be directly inferred from the detailed class II allele calls reported by Otting et al. [25] who relied on full length sequences of these regions in their study (see Supplementary table II). These haplotypes are given preliminary names such as Mamu-DPA02g1 (Table 4).

Table 4.

The 23 common extended MHC founder haplotypes. The column on the extreme right presents the frequency of that particular founder haplotype (R. Wiseman, pers. comm. August 2017; September 2017).

Founder Haplotype #	Mamu-A Haplotype	Mamu-B Haplotype	Mamu- DRB Haplotype	Mamu- DQA Haplotype	Mamu- DQB Haplotype	Mamu- DPA Haplotype	Mamu- DPB Haplotype	Founder haplotype frequency
1	A004	B012b	DR04a	01g2	06_01	02g1	15g1	0.465
2	A004	B012b	DR03a	26g1	18g1	02g1	15g1	0.014
3	A002a	B012a	DR03f	01g1	06g2	07_01	19g	0.085
4	A004	B012a	DR04a	01g2	06_01	02_08	06_04	0.014
5	A008	B028	DR09a	26g2	18g3	06g	01g1	0.014
6	A004	B028′	DR03a	26g1	18g1	02g1	15g1	0.113
7	A008	B047a	DR06	23_01	18_02	02g1	15g1	0.014
8	A001	B047a	DR03a	26g1	18g1	02g2	07_01	0.056
9	A001	B047a	DR09a	26g2	18g3	06g	01g1	0.014
10	A001	B047a	DR04a	23_01	06_01	02g1	15g1	0.014
11	A001	B015b	DR03a	26g1	18g1	02g1	06_02	0.014
12	A004	B001a	DR03a	26g1	18g1	02g2	07_01	0.028
13	A004	B001a	DR04a	01g2	06_01	02g1	15g1	0.014
14	A224a	B001a	DR01a	26g2	18g3	07_01	19g	0.014
15	A004	B048	DR03a	26g1	18g1	02g1	06_02	0.014
16	A004	B048	DR03f	01g1	06g2	06g	01g1	0.014
17	A004	B069a	DR03a	26g1	18g1	02g2	07_01	0.014
18	A002a	B069a	DR04a	01g2	06_01	02g1	15g1	0.014
19	A001	B002	DR03f	01g1	06g2	02g1	15g1	0.014
20	A004	B008	DR06	23_01	18_02	02g1	15g1	0.014
21	A004	B017a	DR01a	26g2	18g3	02g2	07_01	0.014
22	A008	B043a	DR03a	26g1	18g1	02g2	07_01	0.014
23	A002a	B024a	DR02	24g1	18_10	06g	01g1	0.014

As systems of nomenclature for abbreviated haplotypes that define MHC class I and II alleles vary and the data for this study were generated by the WNPRC Genetics Services Unit, their nomenclature system was employed in the present study [23–30]. Three of the seven MHC class I and II regions examined in this study, Mamu-A, -B and –DRB, contain alleles of more than a single locus (see Fig. 1). A combination of alleles of these loci comprises a haplotype of which less inclusive sub-haplotypes that define minor alleles are subsets, and these haplotypes are indicated by a short name/abbreviation for simplification. For example, the haplotype “A008” of the Mamu-A region is defined by its major allele, A1*008 of the A1 locus, and minor alleles of the A2 and A3 loci. Supplementary tables I and II list the abbreviated haplotypes that were identified in each animal by matching sequence reads from the class I and class II regions to previously described class I and class II alleles.

Haplotype diversity, or the number of haplotypes per locus (Ha), observed (HO) and expected (HE) heterozygosity, and FIS were calculated using the program Arlequin version 3 [31]. Deviations from Hardy–Weinberg Equilibrium (HWE) for each of the seven MHC loci were tested using the exact HWE test implemented in GENEPOP 3.2 [32] with the Markov chain method [33] (dememorization: 1000; batches: 500; iterations per batch: 1000). Linkage disequilibrium (LD) among the seven loci, which can be expressed in terms of the correlation coefficient (LD_r ) [34], was tested using the LINKDOS program described by Garnier-Gere and Dillmann [35].

Results

Due to the sensitivity of the MPS technique employed, several thousand (53,000, on average) class I and class II sequence reads were identified in each animal allowing the definition of haplotypes as rare as those that occurred at a frequency of ≤ 1%. The Mamu-A haplotypes are common in Indian origin rhesus macaques with A004, which exhibited a frequency of 57.9% in the cohort studied, being the most predominant haplotype. The Mamu-B haplotypes in this colony are also common Indian origin rhesus haplotypes with B012b being the most prevalent at 36.3%. The DRB haplotypes DR03a and DR04a, respectively, were detected in over 30% of the chromosomes surveyed, and a previously undescribed DRB haplotype, DR-U, was also discovered. Within the DQA/DQB region, the most common DQA haplotypes were 01g2 (34.4%) and 26g1 (31.3%) while the most prevalent DQB haplotypes were 06_01 (34.8%) and 18g1 (31.2%). A majority of DPA haplotypes were 02g1, which occurred among 67.5% of the chromosomes assayed. Among the DPB haplotypes, 15g1 was most frequently detected in the colony (56.9%).

Table 2 provides estimates of genetic diversity parameters for the Mamu-A, Mamu-B, DRB, DQA/DQB and DPA/DPB loci/regions, all of which were relatively high. The number of haplotypes per locus (or per region, in the case of Mamu-A, -B and –DRB [G. Doxiadis, pers. comm. January 2018]) fell between 5 (DPA) and 15 (Mamu-B) with an average of 10.14. High heterozygosity was exhibited for all seven of the surveyed MHC regions. HO estimates ranged from approximately 51% (DPA) to 80% (Mamu-B) with an average of 68.6% while HE estimates ranged from 51% (DPA) to 82% (Mamu-B) with an average of 68.6%. All loci/regions with the exception of Mamu-B (FIS = 0.019) exhibited insignificant levels of inbreeding (FIS = -0.023 to 0.005). While class I and class II haplotype frequencies at all seven loci/regions fit HWE expectations (Table 2), indicating mutation-selection balance; estimates of highly significant LD (p < 0.0001) suggest that the haplotypes within each of these different genomic regions have a strong tendency to co-segregate due to their physical linkage. The wide range of correlation coefficients (LD_r) from 0.099 (Mamu-A-DPB) to 0.465 (DQA-DQB) (Table 3a), however, reveals variation in recombination intensity within the MHC region. For instance, recombination within the chromosomal region between Mamu-A and Mamu-B is relatively more common than that between either of these loci and the DRB region while allelic recombination with the DQA-DQB and DPA-DPB gene pairs occurred at a much lower rate (Table 3b). Figure 1 illustrates the map of the rhesus macaque MHC region based on these LDr estimates.

Table 2.

Genetic diversity estimates across all seven MHC loci/regions. N: sample size, Ha: haplotype number per locus; HO: observed heterozygosity, HE: expected heterozygosity, FIS: inbreeding coefficient and sd: standard deviation. P values denote significant departure from HWE (p > 0.01).

Locus/region	N	Ha	HO	HE	FIS	p value	sd
Mamu-A	379	7	0.610	0.596	−0.023	0.799	0.00027
Mamu-B	379	15	0.799	0.815	0.019	0.611	0.00033
DRB	379	11	0.752	0.752	0.000	0.318	0.00032
DQA	379	10	0.749	0.749	−0.001	0.262	0.00045
DQB	376	8	0.745	0.747	0.003	0.272	0.00041
DPA	379	5	0.507	0.509	0.005	0.365	0.00046
DPB	379	8	0.633	0.635	0.001	0.409	0.00046
Average	378.57	9.14	0.69	0.69	0.00	0.43	0.00

Table 3a.

Analysis of Linkage Disequilibrium between pairs of loci (where r is correlation coefficient, df is degrees of freedom and p is 0.001). Analysis was performed using the LinkDos program [35]

Locus 1	Locus 2	LDr	Χ2	df
Mamu-A	Mamu-B	0.142	1237.67	105
Mamu-A	DRB	0.155	1091.56	77
Mamu-A	DQA	0.158	865.67	70
Mamu-A	DQB	0.110	512.2	56
Mamu-A	DPA	0.108	385.67	35
Mamu-A	DPB	0.099	533.41	56
Mamu-B	DRB	0.148	2080.12	165
Mamu-B	DQA	0.151	1795.01	150
Mamu-B	DQB	0.137	1378.72	120
Mamu-B	DPA	0.109	799.93	75
Mamu-B	DPB	0.101	1110.98	120
DRB	DQA	0.297	3032.04	110
DRB	DQB	0.349	2677.88	88
DRB	DPA	0.155	1102.5	55
DRB	DPB	0.190	1847.95	88
DQA	DQB	0.465	2964.3	80
DQA	DPA	0.142	860.72	50
DQA	DPB	0.193	1669.33	80
DQB	DPA	0.166	850.22	40
DQB	DPB	0.224	1614.54	64
DPA	DPB	0.461	1907.07	40

Table 3b.

Pairwise LDr estimates among the seven loci analyzed.

	Mamu-B	DRB	DQA	DQB	DPA	DPB
Mamu-A	0.142	0.155	0.158	0.110	0.108	0.099
Mamu-B		0.148	0.151	0.137	0.109	0.101
DRB			0.297	0.349	0.155	0.19
DQA				0.465	0.142	0.193
DQB					0.166	0.224
DPA						0.461

Discussion

The distributions and frequencies of MHC class I and class II haplotypes can be examined collectively and comprehensively (Tables 1 and 2) in a large cohort of rhesus macaques. Concordant with previous studies of smaller samples of rhesus macaques by Karl et al.’s [22], Wiseman et al.’s [7, 23] and Otting et al.’s [25], the Mamu-A, Mamu-B, DRB, DQA/DQB and DPA/DPB loci that were analyzed in the present study were all polymorphic with moderate (over 50%) to high (over 80%) levels of heterozygosity. However, our colony-wide assessment of MHC composition revealed that a limited number of Mamu-A, B, and DRB haplotypes accounted for more than half of the MHC variability in the CPRC’s SPF colony. As indicated by the Mamu-class I and class II haplotype frequencies, the MHC composition of this colony, particularly that for the Mamu-A and B haplotypes, is comparable to that of other predominantly Indian origin colonies of the US National Primate Research Centers (NPRC) [23]. For instance, haplotype A002a is common in Indian rhesus macaques and was observed in many individuals in this study [23, 28, 29].

Table 1.

MHC class I and class II haplotype frequencies. See supplementary tables I and II for Mamu-A, Mamu-B and DRB haplotype-specific alleles. For all haplotypes, please refer to https://www.ebi.ac.uk/ipd/mhc/

Mamu-A
A001	A002a	A224a	A004	A057	A008	A023	A031
0.131	0.055	0.009	0.579	0.001	0.222	0.001	0.001
Mamu-B
B001a	B002	B008	B012a	B012b	B015b	B017a	B024a	B028	B028′	B043a	B047a	B048	B055	B069a	B069b
0.095	0.017	0.024	0.05	0.363	0.082	0.005	0.004	0.128	0.011	0.012	0.123	0.044	0.001	0.04	0.043
DRB
DR01a	DR01d	DR02	DR02′	DR03a	DR03f	DR04a	DR06	DR08	DR09a	DR14b	DR-U
0.015	0.001	0.009	0.008	0.313	0.074	0.347	0.111	0.003	0.112	0.001	0.007
DQA
01_05_02	01_06	01g1	01g2	02g1	23_01	23_02	23_03	24g1	26g1	26g2
0.001	0.007	0.067	0.344	0.001	0.113	0.013	0.001	0.004	0.313	0.135
DQB
06_01	06_06	06_09	06g2	18_02	18_10	18g1	18g2	18g3
0.348	0.007	0.001	0.068	0.112	0.004	0.312	0.013	0.134
DPA
02_08	02g1	02g2	04g1	06g	07_01
0.016	0.675	0.127	0.009	0.131	0.042
DPB
01g1	03g	04g	06_02	06_04	07_01	08_01	15g1	19g
0.132	0.009	0.001	0.094	0.015	0.124	0.015	0.569	0.042

Consistent with the study by Doxiadis et al. [36], highly significant linkage disequilibria (p < 0.0001) among all seven loci/regions were observed in this study confirming that the MHC region surveyed tends to behave as a single linkage group. While recombination is low within this region, varying recombination rates across the region have contributed to the rhesus macaque MHC haplotypic structure and diversity. Based on the analyses of the Mamu-A, Mamu-B and DRB regions of a sample of Indian rhesus macaques, Doxiadis et al. [36] argued that MHC haplotype diversity is maintained by recombination events involving a relatively few segments within these loci. However, the overall low rates of recombination and the long persistence times of MHC haplotypes within populations [25, 37] make MHC loci reliable indicators of ancestry and/or geographic origins. For instance, haplotypes Mamu-A 002a and 023 and the Mamu-B-012a, 012b, 017a, 017d, 017f, 024a, 029, 043a, 47a, and 93 are derived haplotypes that are only found in Indian animals while haplotype A002b, which appears to be Chinese in origin, was not observed in this colony [23, 28, 29].

Although rare MHC haplotypes that appear at or below frequencies of 1% in various rhesus macaque breeding colonies have been attributed to admixture between animals from different regional populations [23], none of the rare MHC haplotypes in the present study have been documented to be derived from locations outside the Indian subcontinent. The attribution of the MHC haplotype structure of the CPRC SPF colony solely to its Indian founders is consistent with Kanthaswamy et al.’s [38, 39] studies based on autosomal short tandem repeat (STR) loci which support the sole Indian provenance of these founders. Ancestry assignments based on STRs have demonstrated that the CPRC SPF colony’s degree of Indian ancestry is among the highest of those that have been calculated within primate breeding facilities [38, 39], further confirming the CPRC as a high-value resource for AIDS research [40, 41]. Despite the pure Indian ancestry of the study animals, haplotypes such as Mamu-A 001, 004, 008 and 224a and Mamu-B 001a, 002, 015b, 048 and 069a that were identified in these animals that are shared with Chinese rhesus macaques [23, 28, 29] are regarded as ancestral haplotypes probably due to incomplete lineage sorting of MHC [42].

The extended MHC haplotypes of most CPRC SPF animals (Table 4) were easily inferred mainly due to the extent of LD within this region and the limited number of founders in the source population of Cayo Santiago from which animals were translocated to the CPRC to establish this SPF colony [38, 43]. In addition to being a closed colony, the Cayo Santiago island population, the sole source of the CPRC SPF colony, has experienced extreme founder effects and genetic bottlenecks since its establishment 80 years ago [38, 39, 43, 44]. This is probably why several MHC homozygous individuals with five of the most common Indian-origin haplotypes were identified in this study including 16 individuals that were homozygous for the extended haplotype A004/B012b/DR04a/DQA01g2/DQB06_01/DPA02g1/DPB15g1, which is a haplotype reported to be very common among US primate centers that are predominantly composed of Indian rhesus macaques [R. Wiseman, pers. comm. August 2017; September 2017]. The seven Mamu-A haplotypes observed in this study represent fewer than half of the common Mamu-A haplotypes in Indian-origin rhesus macaques reported in Wiseman et al. [7], reflecting the relatively lower genetic diversity of the CPRC animals. The presence in the CPRC SPF colony of only 18 Mamu-B008 (2.4%) and 4 Mamu-B017 (0.5%) haplotypes (see Table 1 and Supplementary table 1), which have been identified as being critically important for SIV/HIV studies because of their association with exceptional control of SIVmac239 replication during the chronic phase of viral infection, also underscores the relative isolation of this colony compared to other breeding colonies in the US [13].

MHC typing of the CPRC SPF colony is critical for selecting animals for specific experimental studies and the management of this colony. While the results reveal that this relatively genetically homogeneous colony would be a good source for MHC haplotype-matched groups of animals for investigators, the array of MHC haplotypes exhibited within it still allows a significant likelihood for MHC incompatibility among study animals. In either case, information on the composition and distribution of MHC class I and class II haplotypes within the CPRC SPF colony will enhance the investigator’s ability to predict and understand immune responses mounted by the majority of these SPF animals and prevent risks of incompatibility.

As knowledge of the MHC genetic composition of colony animals is also fundamental for making sound colony management decisions, lacking this data will make it impossible to anticipate and meet investigator requirements for particular types of MHC haplotypes. For instance, in the past, selection regimes at the CPRC were designed to increase the percentage of Mamu-A*001-positive animals in the colony. While such schemes for selective breeding can increase the proportion of animals in a colony that is homozygous for Mamu-A*001, they are ineffective for increasing Mamu-A*001 allele frequencies unless the reproductive success of Mamu-A*001-positive breeders can be increased above that of Mamu-A*001-negative breeders. Nevertheless, more recently, HIV/AIDS researchers have more often routinely requested Mamu-A001 negative animals instead. In order to avoid any dramatic fluctuation in colony-wide frequency of MHC haplotypes or unintentional elimination of valuable haplotypes from the colony, as one haplotype is favored over another by researchers, colony managers at the CPRC are now choosing male breeders with common MHC haplotypes and female breeders with rare MHC haplotypes. In the short term, a breeding scheme that prioritizes the maintenance of rare alleles or haplotypes over heterozygosity could result in substantially depleting diversity at the MHC as well as other loci in the genome [45]. To ameliorate any potential long-term losses in MHC variability, novel MHC variation can be introduced into the colony by translocating individuals from Cayo Santiago and/or other genetically and demographically healthy NPRC colonies.

Beyond the characterization of class I and class II haplotypes, information about the larger-scale genetic structure of the macaque MHC region and diversity of MHC genes other than those within class I and class II regions remains somewhat restricted. Advances in the application of cost-effective, high throughput and deep sequencing technologies have triggered comprehensive MHC testing of large macaque cohorts [7]. Continued improvements in the detection methods for MHC typing, including enhanced next-generation techniques, will facilitate a better understanding of the immune response, not only to SIV infection but infection by many other pathogens as well, and expand the usefulness of rhesus macaques in biomedical research. However, with the availability of more MHC information, colony managers must be cautioned that breeding and animal selection schemes designed strictly around this single linkage group will risk losing variability at other potentially important loci [46–48]. Overall genetic diversity can be conserved by employing colony management protocols that rely on strategies to monitor genetic diversity across a panel of unlinked autosomal loci that is distributed throughout the genome such as STRs or SNPs [38, 39].