Haplotype analysis of the mitochondrial DNA d‐loop region reveals the maternal origin and historical dynamics among the indigenous goat populations in east and west of the Democratic Republic of Congo

Abstract This study aimed at assessing haplotype diversity and population dynamics of three Congolese indigenous goat populations that included Kasai goat (KG), small goat (SG), and dwarf goat (DG) of the Democratic Republic of Congo (DRC). The 1169 bp d‐loop region of mitochondrial DNA (mtDNA) was sequenced for 339 Congolese indigenous goats. The total length of sequences was used to generate the haplotypes and evaluate their diversities, whereas the hypervariable region (HVI, 453 bp) was analyzed to define the maternal variation and the demographic dynamic. A total of 568 segregating sites that generated 192 haplotypes were observed from the entire d‐loop region (1169 bp d‐loop). Phylogenetic analyses using reference haplotypes from the six globally defined goat mtDNA haplogroups showed that all the three Congolese indigenous goat populations studied clustered into the dominant haplogroup A, as revealed by the neighbor‐joining (NJ) tree and median‐joining (MJ) network. Nine haplotypes were shared between the studied goats and goat populations from Pakistan (1 haplotype), Kenya, Ethiopia and Algeria (1 haplotype), Zimbabwe (1 haplotype), Cameroon (3 haplotypes), and Mozambique (3 haplotypes). The population pairwise analysis (FST ) indicated a weak differentiation between the Congolese indigenous goat populations. Negative and significant (p‐value <.05) values for Fu's Fs (−20.418) and Tajima's (−2.189) tests showed the expansion in the history of the three Congolese indigenous goat populations. These results suggest a weak differentiation and a single maternal origin for the studied goats. This information will contribute to the improvement of the management strategies and long‐term conservation of indigenous goats in DRC.


| INTRODUC TI ON
In most developing countries, agriculture and particularly livestock farming constitute an important source of income in rural households (Herrero et al., 2013). In that respect, goats, one of the first domesticated animals, provide meat and milk as a major source of income for smallholder farmers (Aziz, 2010;Baenyi Simon et al., 2021;Naderi et al., 2008;Skapetas & Bampidis, 2016).
Estimated to 4,065,709 heads, indigenous goat populations in DRC are grouped into three local breeds locally called "chèvre moyenne du Congo" or small goat (SG) of Congo, "chèvre du Kasai" or Kasai goat (KG), and "chèvre de Bandundu" or dwarf goat (DG) of Congo (FAOSTAT, 2015;Lafleur et al., 2018;accessed March, 2021). These goats are spread throughout all the agro-ecological zones (AEZs) of the country where they are kept by farmers (FAOSTAT, 2015;accessed March 2021). In South Kivu, goats contribute up to 40% to farmers' household income (Baenyi Simon et al., 2021;Wasso et al., 2018). As in eastern Africa, goats in DRC are raised in marginal areas, where crops production is not possible, in different production systems with the predominance of an extensive system that is characterized by low breeding inputs (Muigai et al., 2017).
Due to uncontrolled livestock movements across borders, exotic goat breeds would have been introduced and crossbred with the three Congolese indigenous goat populations. Such a practice increases the risk of the disappearance of resilient and adapted local breeds. To mitigate this risk, the locally adapted goat breeds in DRC need to be characterized, conserved, and utilized sustainably. Characterization of local breeds provides large knowledge and gives a clear perspective on the population structure that will assist in the decision-making of future breeding programs (Groeneveld et al., 2010;Yang et al., 2016). Because animal mitochondrial DNA (mtDNA) evolves faster than a nuclear genetic marker, it represents a good informative region for the study of phylogenetic and evolutionary biology (Ladoukakis & Zouros, 2017). It also permits the faster examination of the relatedness of populations and has become important in biogeographic and anthropologic studies (Lehman & Fleagle, 2006). The mtDNA polymorphism, especially the displacement loop (d-loop) region, is one of the important tools that have been used to better understand the genetic diversity, the population structure, and the population dynamics in different animal species including goats (Phyu et al., 2017;Tarekegn et al., 2018Tarekegn et al., , 2019, sheep (Agaviezor et al., 2012), and chickens (Liu et al., 2009).
In general, six mtDNA haplogroups (A, B (B1 and B2), C, D, F, and G) were identified and found distributed in different geographic areas in the world. The haplogroup A was shown to have a large geographic distribution (Pereira et al., 2005) and was more reported in a large part of African regions (Luikart et al., 2001). However, haplogroups B and G were also reported in some African countries; with haplogroup B particularly found limited in the South part of Africa, especially in South Africa and Namibia, while the haplogroup G was reported in Egypt (Naderi et al., 2007), Kenya (Kibegwa et al., 2016), Ethiopia (Tarekegn et al., 2018), Sudan, andSomalia (Al-Araimi et al., 2017).
A previous study on indigenous goat from Peste des Petits Ruminants outbreak zones in South Kivu province of DRC based on mtDNA d-loop variation revealed the presence of two haplogroups A (commonest) and B within the goat population in South Kivu (Bwihangane et al., 2018). However, the result from this study was limited to goats from Peste des Petits Ruminants outbreak zones in the South Kivu region and could not reveal more information on the genetic diversity of indigenous goat breeds in the whole country and did not mention the goat populations' dynamics and history. Therefore, this study aimed to describe the haplotype diversity, the population structure, and the demographic dynamics of three indigenous goat breeds in three AEZs of DRC based on the mtDNA d-loop region.

| Sampling and DNA extraction
Sampling was conducted in collaboration with the Ministry of Agricultural, Livestock, and Fisheries of DRC through the representative inspections in each sampling region that included Kinshasa, Tshopo, and South Kivu (Figure 1). populations. Negative and significant (p-value <.05) values for ) and Tajima's (−2.189) tests showed the expansion in the history of the three Congolese indigenous goat populations. These results suggest a weak differentiation and a single maternal origin for the studied goats. This information will contribute to the improvement of the management strategies and long-term conservation of indigenous goats in DRC.
A total of 339 blood samples representing three Congolese indigenous goat populations (Kasai goat, n = 108; dwarf goat, n = 114 and small goat, n = 117) were sampled from farmer's flocks from the three AEZs of DRC (representing in this study by Kinshasa, Tshopo and South Kivu) and used for the study. Efforts were made to avoid closely related individuals during sampling. The socioeconomic factors associated with goat keeping (Wasso et al., 2018), the environmental characteristics (the high land volcanic mountain), and the proximity to neighboring countries (Tanzania, Rwanda, and

| PCR amplification and sequencing of the dloop region
The 1169 bp of the mtDNA d-loop region was amplified using primers designed (F: 5'-ACCAGAAAAGGAGAATAGCC-3'; R: 5'-TGATCTAGTGGACGGGATAC-3) were respectively used as external and internal primers to sequence the purified PCR products .

| Data analysis
Default values and parameters inherent in algorithms and software were used for all analyses undertaken in this study. Only deviations from the default were mentioned. Before the analyses, all the chromatograms were visualized using CLC Genomics workbench v8.0 software. MEGA v6.4 software was used for the multiple sequence alignments with the ClustalW algorithm (Tamura et al., 2013). The variable sites were scored against the Capra hircus reference sequence (GenBank accession number: GU223571: direct submission). In total, 339 sequences were generated from where haplotypes were generated with DnaSP v5 (Rozas et al., 2003).
Genetic diversity parameters that include the number of haplotypes (N), haplotype diversity (H d ), nucleotide diversity (π), and mean of nucleotide differences between haplotypes (K) and their standard deviations (SD) were analyzed for each goat population and across all populations using DnaSP v5 (Librado & Rozas, 2009;Rozas et al., 2003).
A phylogenetic tree was constructed using haplotypes generated in Congolese indigenous goat and 22 reference haplotypes representing the 6 haplogroups (A, B, C, D, G, and F) defined based on the variation in the first HVI with 481 bp of length size (Luikart et al., 2001;Naderi et al., 2007Naderi et al., , 2008  haplogroups (Naderi et al., 2007) were retrieved from the GenBank (Table S1) and included in the NJ tree and MJ network analyses using Network v4.6 software (Bandelt et al., 1999). The added sequences were from 13 African countries: Cameroon (central Africa), Kenya, Ethiopia (East Africa), Egypt, Algeria, Libya, Tunisia and Morocco (North Africa), Senegal, Nigeria (West Africa), Namibia, Zimbabwe, and Mozambique (southern Africa). In addition, sequences from 10 Asian countries (India, Iraq, Saudi Arabia, Pakistan, China, Laos, Iran, Mongolia, Jordan, and Azerbaijan) and 6 European countries (Turkey, Austria, France, Italy, Switzerland, and Spain) were included in the analysis.
The genetic variation among Congolese indigenous goat populations was evaluated through the analysis of molecular variance (AMOVA) following 1000 permutations in Arlequin v3.5.2 (Excoffier & Lischer, 2010). Pairwise genetic differentiations (F ST ) (Reynolds et al., 1983) were estimated between each of the three Congolese indigenous population and a group consisting of the non-Congolese goat populations grouped according to the geographical regions using Arlequin v3.5.2 software (Excoffier & Lischer, 2010) with the number of permutations for significance estimated at 100 at the significance level of 0.05.
We inferred population demographic history and dynamics from haplotype mismatch distribution patterns (Rogers & Harpending, 1992) and the expected distributions plus their 95% confidence intervals for the three Congolese indigenous goat populations.
Departures of the observed sum of squares differences (SSD) from the simulated model of expansion were tested with the chi-square test of goodness of fit statistic and Harpending's raggedness index "r" (Harpending, 1994) following 1000 coalescent simulations.
Analysis of mismatch distribution patterns was augmented with two coalescent-based estimators of neutrality: Fu's Fs (Fu, 1997) and Tajima's D (Tajima, 1989) statistics. The significance of these two statistics was tested with 1000 coalescent simulations in Arlequin v3.5.2. Outside Africa, one haplotype was shared with Pakistan goat.

| Population phylogenetic and relationship analysis
The first hypervariable (HVI, 453bp) sequences both for Congolese indigenous goat populations and the reference sequences representing the six domestic goat haplogroups were aligned using the ClustalW algorithm implemented in MEGA v6.4 and considered for the NJ tree analysis. All the three Congolese indigenous goat populations (100%) were clustered into mtDNA lineage A (Figure 2). An MJ (as described in the data analysis section) constructed to provide a wider resolution of the phylogenetic relationship between the Congolese indigenous goat populations and the non-Congolese goats (Figure 3) supported the result obtained by the NJ tree analysis.

| Population genetic structure
The AMOVA for the three Congolese indigenous goat populations grouped by AEZs revealed that 5.88% of the total genetic variation was attributed to the genetic differences among populations and the highest proportion (94.12%) to variation within populations ( Table 2).  (Table 4).

| Population dynamics and history
Mismatch distribution patterns were used to assess the population dynamics of the three Congolese indigenous goat populations grouped into the haplogroup A as revealed by the NJ tree and the MJ network. For each population, the expansion modal of mismatch distribution was bimodal and a valid goodness of fit was observed between observed and expected distributions plus their 95% confidence intervals, indicating a demographic expansion signal (Figure 4).
For the global dataset incorporating the three Congolese indigenous goat populations, the observed patterns of mismatches did not deviate significantly from that expected under a null hypothesis either with the demographic (p-value >.05) or spatial (p-value >.05) expansion modal (Table 5). Exception was observed for the dwarf and small goats for which the SSD values were significant (p-value <.05).
The variations around the curves were also not significant except for dwarf goats (p-value <.01) if considering the demographic expansion modal (Table 5)  geographical locations suggest that the gene flow has occurred on a regional scale during some time in the recent past and the breeds have not been subdivided by long-term biogeographic barriers (Luikart & Allendorf, 1996). In a study conducted by Vacca et al. (2010), it has been reported that goats have shown high genetic haplotype diversity, from where more haplotypes were each represented by a single sequence and only a few were shared among animals. Results in this study support that finding and the findings of Naderi et al. (2007) showing that it is common to find haplotypes represented by one individual or shared by only a few individuals. Based on the available goat mitochondrial haplogroup classification system (Naderi et al., 2007)   The haplogroup A has been shown to have a large geographic distribution (Pereira et al., 2005) and was more reported in large parts of African regions (Luikart et al., 2001). That result could be interpreted as evidence that Congolese indigenous goats come from a unique maternal population with one maternal lineage which could have been introduced from one geographic domestication area (Naderi et al., 2008). On the one hand, the fact that 3 and 1 haplotypes were shared between Congolese and Mozambican, and Congolese and  (Moritz et al., 1987;Tserenbataa et al., 2004) and relates to high haplotype diversity implying widespread distribution and diversity to favor for selection within populations (Kibegwa et al., 2016). A weak phylogeography with small genetic differentiation (F ST ) was confirmed between the three studied Congolese indigenous goat populations (Table 3). This low genetic differentiation points to a high historical gene flow and intermixing between the three goat populations studied. Consistency is observed between the result in this study and previous reports confirming that weak genetic structure is most observed in small ruminants (sheep and goat) than in large ruminants (cattle) (Luikart et al., 2001). Thus, findings in this study suggest that the three Congolese indigenous goat populations share a relatively similar genetic background with the same maternal origin as revealed by the demographic dynamics inferred from the mismatch distribution pattern (Figure 4). The mismatch distribution patterns were bimodal for the three Congolese indigenous goat populations. Similar demographic patterns have been observed in Tanzanian (Nguluma et al., 2021), Ethiopian ,  Swedish International Development Cooperation Agency (Sida).

Percentage of variation
The University of Nairobi, UoN also provided initial support on project writing.

CO N FLI C T O F I NTE R E S T
No conflict of interest was reported by the author(s).

DATA AVA I L A B I L I T Y S TAT E M E N T
Mitochondrial sequence data generated as part of this project are deposited in dryad and given accession number https://doi. org/10.5061/dryad.vq83b k3v6. Other mtDNA sequences incorporated into the analysis were downloaded from this source and can be retrieved as per the relevant citations.