Genetic Diversity Analysis among 27 Indian Goat Populations using Microsatellite Markers

Genetic diversity studies of livestock populations mainly focus within- and between-breed diversity, breed history and their adaptive variations. It also assesses genetic uniformity, admixture or subdivision, inbreeding or introgression in the population which is helpful in breed formation and their sustainable utilization. Conservation and improvement strategies are based on proper genetic characterization in association with phenotypic evaluation (Tadelle, 2003; Halima, 2007).
 
India has 34 (as per the latest figure of breeds registered with NBAGR) goat along with many non-descript goat populations which are widely distributed and adapted to different agro-climatic regions of the county. However, dilution of these genetic resources has occurred due to intermixing, sub-structuring and consequent genetic drift in the population over the time. Therefore, analysis of genetic variation provides genetic information that would be helpful in maintenance and management plans, formulating appropriate mating systems and reducing inbreeding rate and consequently deciding the conservation priority to maximize the diversity and to make guidelines for a constant breed improvement strategy for meat and milk production. The genetic diversity analyses of 22 different goat populations were published by Dixit et al., (2009). Here, we have included five more non-descript new goat populations.
 
Among Narayanpatna, Raighar, Kalahandi and Malkangiri  are non-descript goats  found in Narayanapatna block of Koraput district, Raighar and Umarkote block of Nabarangpur district, Kalahandi district and Malkangiri district of Odisha state of India (Kornel et al., 2006) respectively. These populations show early maturity, are highly prolific and mostly kids twin. Although, there is no evidence regarding origin of these population but Gollas, Goudas or Yadvas and ethnic tribal has maintained them over long years. Koya and Matia tribes are believed to be responsible for development of Malkangiri goat; Bhattaras and Gond for Raighar and Kondhs for Narayanpatna with their special interest due to their socio-religious, economic use and dietary habits. These goats are mostly maintained under extensive and rarely under semi-extensive systems and usually 1:5 or 1: 10 male and female ratio maintained. In coastal region, there is high mortality in these populations. Usually male is hired on payment basis for breeding purposes (Kornel et al., 2006). Phenotypic and biometric studies of Odisha goat populations (Verma et al., 2015) showed considerable variation. The specific objective of this study was to identify and classify these non-descript goat populations into separate identity group(s) and diversity analysis of all these resources using microsatellite markers. Another non- descript goat population included in the study was Rohilkhandi from Rohilkhand region of Uttarakhand state.

Data generation
 
The present research work was conducted during the year 2016-17 at National Bureau of Animal Genetics Resources, Karnal-132001. STR data of 25 markers on 1237 random samples of 27 goat populations was used for analysis. The allelic data was generated using 25 microsatellite (http://cabin.iasri.res.in/gomi/algorithm.html) loci based on DNA fingerprinting of 27 goat populations. PCR products were mixed and resulting mixture was denatured by incubation to run on automated DNA sequencer of Applied Biosystems. The electropherograms were drawn through Gene Scan and used to extract DNA fragment sizing details using Genemapper software 4.0. Generated data is numeric in terms of base pair which is size of each allele along with genotype (combination of allele at every diploid locus). The protocol followed was same as described at http://cabin.iasri.res.in/gomi/tutorial.html. 
 
Data analysis
 
Genetic diversity analysis was carried out to determine the genetic variation within and between breeds/populations; parameters such as heterozygosity and Wright’s F-statistics (F_ST, F_IS and F_IT). Heterozygosity was defined as the probability that a given individual randomly selected from a population will be heterozygous at a given locus. The F_ST is an estimate of variation due to differences among populations which is the reduction in heterozygosity of a subpopulation due to genetic drift. The F_IS is an estimate of variation within populations that measures the reduction in heterozygosity in an individual due to non-random matting within its population. The F_IT is the overall inbreeding coefficient of an individual relative to the total population.
       
Genetic diversity was assessed by calculating the observed and effective number of alleles (N_A and N_E), mean number of alleles (MN_A), observed heterozygosity (H_O) and expected heterozygosity (H_E), Nei’s genetic distance and pairwise F_ST by using GENALEX version 6.5 (Peakall and Smouse, 2012). Polymorphism information content (PIC) and F_IS was calculated by using Molkin version 3.0 (Gutierrez et al., 2009).
 
Structuring of goat populations
 
The genetic structuring of the sampled population was carried out by clustering techniques using a Bayesian clustering procedure with the admixture method implemented in STRUCTURE software ver. 2.3.4 (Pritchard et al., 2000), (http://pritch.bsd.uchicago. edu/structure.html). The program can estimate “hidden structure” i.e. the number of different clusters (K partitions) obtained without using any prior information about individual membership (population and/or breed). Moreover, for each individual, the program is able to determine the proportion of genes originating from the K potential clusters. Based on the Markov Chain Monte Carlo (MCMC) method, the STRUCTURE algorithm uses a model-based approach, under the hypothesis that the loci are in Hardy-Weinberg equilibrium and in linkage equilibrium within clusters, to define the natural logarithm of the probability that a given genotype belongs to the assumed K clusters. The clustering of 27 studied populations was analysed by using independent allele frequencies model. 50 runs for each value of K (2≤K≤ 27) was done with 60,000 interactions following a burn-in period of 100,000. Pair-wise comparisons of the 50 solution of each K value were run along with 50 permutations. The most probable clustering numbers (best ΔK value) was assessed according to the equation: ΔK = m (|L” (K)|)/ s[L (K)], (Evanno et al., 2005).
 
A phylogenetic tree was constructed based on the Nei’s genetic distance (D_A) by using neighbour-joining (NJ) method (Saitou and Nei, 1987). The robustness of the topologies was evaluated with a bootstrap test of 1000 resampling across loci. Phylip ver 3.695 software programme by Joseph Felsenstein (2013) was used to construct the phylogenetic tree of population relationships. Finally ANOVA analysis considering groups and populations as source of variance was assessed by ARLEQUIN version 3.5 (Excoffier and Lischer, 2010).
 
Principal Component Analysis (PCA)
 
In India, goat rearing practices is mainly community based and bucks are used for breeding purposes mainly on the dam performance. Therefore, a multivariate procedure was used to represent population relationship. Multivariate procedure is recommended because the admixture populations are not original evolutionary units and may be misrepresented by NJ phylogenetic tree (Cavalli-sforza et al., 1994). If a compromise structure exists, then a multivariate analysis will be meaningful. The PCA was done using allelic frequencies as a variable after congruent test.

Microsatellite marker’s polymorphism
 
The observed number of alleles ranged from 4.73 (ILSTS005) to 15.0 (OarFCB304) whereas expected number varied from 2.13 (LST008) to 7.71 (OarFCB304) with mean value of 3.85± 0.07(Table 1). The mean number of observed alleles per locus was found to be higher than expected which indicated immigration of alleles in these goats. The overall allelic diversity is considered to be a reasonable indicator of genetic variation. A microsatellite preferably should have at least 4 alleles to be useful for the evaluation of genetic diversity as per the standard selection of microsatellites loci (Barker, 1994).
 

 
The polymorphic information content (PIC) of a marker reveals its usefulness in diversity analysis of a breed. Following the criteria of Botstein et al., (1980), 84% of the investigated markers were observed to be highly informative (PIC > 0.5), 12% as reasonably informative (0.25 < PIC < 0.5) and only 4% were slightly informative (PIC < 0.25). PIC value ranged from 0.607 (ILSTS022) to 0.933 (OarFCB304) with an average value of 0.794 (Table 1), which again indicated abundant genetic diversity in the population. The higher PIC further indicated the utility of these markers for population assignment (Mac-Hugh et al., 1997) as well as genome mapping (Kayang et al., 2002) studies in addition to genetic diversity analysis.
 
Genetic variability is also measured as the amount of actual or potential heterozygosity, as presented in Table 1. Expected heterozygosity was found to be higher than the observed heterozygosity at all the loci. The mean observed and expected heterozygosities were 0.55± 0.01 and 0.67± 0.01, respectively. Most of the loci showed relatively higher expected heterozygosity values that might be due to low selection pressure, large population size and immigration of new genetic material. Thus, there was a considerable genetic polymorphism within and between populations on the basis of their allele number per locus (NA) and their genetic heterozygosity (H_E). The usefulness of these markers in diversity analysis was also indicated in the earlier study of Dixit et al., (2009).
 
Overall means of F_IT, F_ST and F_IS obtained from Jackknifing over loci were significantly different from zero. The F_ST value ranged from 0.084 (OarFCB304) to 0.382 (ETH225) with an overall genetic differentiation of 18.3% among breeds which indicating that there is genetic differentiation among the studied population and remaining 81.8% corresponding to differences among individuals within population. The F_IS value ranged from -0.010 (ILSTS059) to 0.641 (OarJMP29). An overall significant heterozygote deficit (F_IS) of 18 % was observed over all loci within samples. The positive F_IS was resulted from genetic subdivision, non- random mating. The heterozygote deficit and moderate genetic differentiation among 22 goat populations of India were also reported by Dixit et a., (2009). Moreover, F_IS is used to show degree of inbreeding and endangerment potentiality and considered an important tool to judge the conservation priority (Simon et al., 1993). Accordingly, when F_IS< 0.05 then breed is considered as not endangered, 0.05 to 0.15 is potentially endangered, 0.15 to 0.25 is minimally endangered, 0.25 to 0.40 is endangered and > 0.40 is considered as critically endangered.
 
Molecular genetics diversity of population
 
The population genetic diversity and genetic distance of 27 goat populations over 25 STR markers are presented in Table 2. The mean number of alleles across the loci was higher than 8 in more than half of the populations. The mean number of alleles ranged from 5.640±0.661 (Osmanabadi) to 9.360±0.519 (KanniAdu) followed by 8.920±0.700 in Narayanpatna. A high MNA indicated the presence of great genetic variation which may be due to cross breeding or admixture. While the low value as in case of Osmanabadi and Malkangiri indicated low variation due to genetic isolation, historical population bottleneckor founder effect. Rout et al., (2008) also estimated mean number of alleles in the range of 8.1 (Barbari) to 9.7 (Jakhrana) in Indian goats.
 

 
An appropriate measure of genetic variation was gene diversity (average expected heterozygisity). Among the breeds, the expected heterozygosity was found to be higher than observed heterozygosity. The average value of observed heterozygosity was 0.549±0.010 with a range of 0.421±0.057 in Osmanabadi to 0.643±0.053 in Mehsana whereas the average value of expected heterozygosity was 0.669±0.007 with a range of 0.566±0.048 in Osmanabadi to 0.731±0.021 in KanniAdu. The similar values of gene diversity were also reported in literature (Rout et al., 2008; Dixit et al., 2009 and Serrano et al., 2009 />  
The F_IS value ranged from 0.046 (Malkangiri) to 0.335 (Jharkhand Black) which indicated that Jharkhand Black had more deviation from Hardy–Weinberg Equilibrium. F_IS value of non-descript goat populations were on lower side of average F_IS value (0.192) of all the populations. This reflected random mating in these populations. The moderate value of F_IS indicated some of the loci in the breeds were homozygous presumably resulting from mating between relatives and consequent genetic drift which also reflect degree of endanger. Same level of inbreeding was also reported in different goat population (Dixit et al., 2009 and Vahidi et al., 2014).
 
The pair-wise Nei’s genetic distance (D_A) and F_ST statistic values among 27 Indian goat populations are presented in Table 3. The low genetic distance was indicative of closeness and vice-versa. The high Nei genetic distance (>0.5) was observed among most of the Indian registered goat breeds except few. Highest (1.00) genetic distances were observed between non-descript studied new population and many of registered breeds like Black Bengal, Ganjam, Gohilwadi, Jharkhand Black, Attapaddy, Changthangi, Kutchi, Mehsana, Sirohi and Malabari while in some cases, it was low. Though, among non-descript populations, there was low genetic distance (< 0.24) but Narayanpatna was more distantly placed with rest of the non-descript populations, followed by Raighar which is almost same as represented by average co-ancestry (Table 6). The lowest pair-wise Nei’s genetic distance was between Jamunapari and Marwari followed by Raighar and Rohilkhandi (0.09) and highest genetic distance (1.00) was also observed between some of breeds. Rout et al., (2008) estimated lowest genetic distance between Marwari and Sirohi as 0.135 and highest (0.246) between Pashmina and Black Bengal. The estimates of Mohmoudi et al., (2011) ranged from 0.273 to 0.745 among Iranian goat populations. The  lower estimates of F_ST value in range of 0.03 to 0.08 was found among Black Bengal, Ganjam, Gohilwadi, Jharkhand Black, Attapady, Jakrana, Surti, Gaddi, Marwari, Barbari, Beetal and KanniAdu while the highest FST value was estimated between non- descript studied population and Black Bengal, Ganjam, Gohilwadi, Jharkhand Black, Attapady, Changthangi, Kutchi, Mehsana, Sirohi and Malabari. The F_ST value among new populations was in the range of 0.02 to 0.05. Thus, the study revealed F_ST value among new populations and high Nei’s genetic distance of new goat populations with other studied goat breeds. Hence, new populations were not significantly different from each other but with rest of breeds.
 

 
The analysis of molecular variance (AMOVA) within and between populations is presented in Table 4. The AMOVA revealed that percentage of variation among population was 28.18 and within populations were 71.82. Rout et al., (2008) estimated variance among Indian goats as 6.59% which was lower than present findings probably due to fewer numbers of breeds considered. Zaman and Chandra Shekhar (2015) studied the genetic diversity and population structure of four goat populations of Northeast India including West Bengal and showed 21% of the total variation was due to differences between genetic groups.
 

 
Population structure
 
The structure and clustering software have ability of inferring the correct number of subpopulation and assigning individuals appropriately even when genetic differentiation among groups is low (0.02 to 0.05) (Latch et al., 2006). In present study, 1237 individuals from 27 populations sub-clustered by STRUCTURE are presented in Fig 1 and most of the populations reached their own distinct cluster containing only a single population. The results derived from use of this programme provide a strong support of new population cluster subdivision. This subdivision seems to be reasonable since few farmers in studied areas exchange goats and therefore these population show more genetic homozygosity. Ganjam, Gohilwadiand Malabari fall in one group; Jamunapari, Jakhrana, Marwai, Barbari and Beetal in 2^nd group; Sangamneri, Changthangi and Osmanabadi in separate group and rest of the breed fall under their own group. Among non-descript studied new populations, Narayanpatna and Raighar clustering in one group while Kalahandi, Rohilkhandi and Malkangiri in another cluster. A different approach is that individuals belonging to different clusters could be used in planned mating to maintain a good level of genetic variability and rusticity (stress-resistance) and avoid excessive inbreeding (Guastella et al., 2010).
 

 
Phylogenetic analysis
 
Takezaki and Nei, (1996) have demonstrated that D_A (distance) and D_C (diversity) are the most efficient means of obtaining a correct topology on the basis of microsatellite analysis when within population variation is high and distance between each pair of population are used to build a tree. The phylogenetic tree analysis is presented in Fig 2. which revealed that the smallest distance is between Jamunapari and Marwari followed by Kalahandi and Rohilkhandi while largest distances are between B.Bengal with Jamunapari, Marwari, Barbari; Ganjam withJamunapari, Marwari; Gohilwari with Jamunapari, Marwari, Jakhrana, Barbari Zalawadi; Jh. Black with Jamunapari, Marwari, Barbari and New populations with B. Black, Ganjam, Gohilwari, Jh.Black, Attapady, Changthangi, Kutchi, Mehsana, Sirohi and Malabari. The smallest distance is seen among new goat populations compared with other goat breeds which are clearly visible with considerable reliability in phylogenetic tree. The phylogenetic study revealed that Rohilkhandi populations are separated from Malkangiri, Raighar and Kalahandi with 70% reliability and again Narayanpatna population from rest of the new population with 95% reliability. These new population are meat type. Except few, almost all known Indian milk breed clustering in one cluster and separated with new population with 57.5% reliability. Again milk and meat breed/populations are separated with rest of the breeds with 94.6% reliability.  The present studied breed/population also shows that almost all breeds/ populations are clustered according to their genetic distances but not all; some are according to geographical distance which indicates limited exchange of animals may be due to socio-cultural constraints. Serrano et al., (2009) study the genetic structure using Bayesian clustering of 22 Guadarrama goats and found that there is no correlation between geographical distances and genetic distances regarding distribution of breeds. In greek sheep, Ligda et al., (2009) shown that the phylogenetic relationships are in accordance with the geographical location of the breeds, the history of the origin of the breeds and the breeding practices. Mahmoudi et al., (2010) analyzed genetic distance using an unweighted pair group method with arithmetic means (UPGMA) diagram based on Nei’s standard genetic distances, yielded relationships between populations was agreed with their origin, history and geographical distribution. Hassen et al., (2012) grouped the six goat populations with Neighbor-joining using UPGMA methods with bootstrap value of 1,000 into two major groups viz. Agew, Gumuz, Bati, Begia-Medir as first group and Central Abergelle goats as the second group. They obtained higher total variation within the goat populations (95%) confirms a close relatedness of the studied goat ecotypes, which might have happened due to existence of uncontrolled breeding resulting from movement of animals through various market routes and agricultural extension systems.
 

 
Principal component analysis
 
The principal component analysis is presented in Fig 3. which clearly revealed three clusters of the breeds. The first cluster consisted of new populations and Zalawadi belonging to nearby coastal area, second cluster consisted of breeds of western costal region of north-western region, southern-peninsular India and those of eastern region; and third cluster consisted of breeds of northern (gangetic plain) and western plains of north-west region including lower Himalayan region, However, Attappady and Kanniadu of Peninsular India also joins the third cluster. Malabari and Zalawadi were distinct from other breeds. Thus, PCA analyses clearly separated the breeds of Himalayan regions from rest of the breeds. Thus, there seems to be three major point of evolution of Indian goat breeds based on microsatellite markers. On the other hand, if we see on their functionality; they are clearly separated in three clusters except few exceptions.  New population along with Zalawadi is grouping in one group as they are sharing coastal area. Most of the milk breeds viz- Jamunapari, Barbari, Beetal Marwari, Surti, Jakhrana belonging to plain area are grouping in one group while the rest of the breeds are in another group. In general, milch and dual purpose goats are clustering in one cluster while meat purpose breeds in another cluster.
 

 
Thus, investigation of the phylogeny and the plot for PCA indicated that Indian goat breeds were grouped according to their physio-geographic location. Rout et al., (2008) reported that both phylogenetic tree and PCA showed the distribution of Indian goat breed basically in two major clusters with respect to geographical distribution
It also seems that there is a clear association between particular goat type and their sub-region. Within sub-region, populations could be grouped according to phenotypic characteristics except few.
 
Average co-ancestry among subpopulations
 
The molecular co-ancestry information is a useful tool to study the genetic relationship between the breeds. Both average kinship distance (D_k) and average molecular co-ancestry coefficient (f_ij) account for the allele frequencies in the founder population whereas Nei’s genetic distance and Reynold’s genetic distance characterize the short term evolution of the population (Alvarez et al., 2005). The molecular co-ancestry based parameters may be used with classical genetic parameters to obtain the information on population dynamics in livestock as suggested by Alvarez et al., (2005). In present study, the average co-ancestry among subpopulation i and j, i.e  f_ij average kinship distance and Nei’s minimum genetic distance were estimated using allelic frequency over 25 microsatellite loci between 27 studied goat breeds/populations and are presented in Table 5 and Table 6. The average co-ancestry within population ranged from 0.167 (Zalawadi) to 0.409 (Osmanabadi) but most of the populations (19) had average co-ancestry >0.300 which indicated a strong co-ancestry within individuals of a population. But between populations, it was low in most cases. The average kinship distance within breeds was around 0.4 while between breeds, it ranged from 0.418 (Black Bengal and Malabari) to 0.642 (Jharkhand black and Marwari). The new populations under study showed >0.5 kinship distance with rest of breeds except Osmanabadi and Zalawadi (≤ 0.4). The Nei’s minimum distance, value was very lower than kinship distance but the trend was almost same except few cases. Traore et al., (2009) also computed molecular co-ancestry value between goat breeds which ranged from 0.418 to 0.450. The new goat populations had very low average co-ancestry relationship with Black Bengal, Ganjam, Gohilwadi, Jharkhand Black, Attapady, Changthangi, Kutchi, Mehsana, Sirohi and Malabari, which was in agreement with high F_ST value estimated among them, but it was high with rest of the breeds.
 

 

The present work for the first time generated information about how new goat populations are related with existing Indian goat populations and among themselves based on microsatellite analysis. They have a unique gene pool with good adaptation to the coastal region of the country. Genetic parameters showed a high value of inbreeding and therefore should be monitored for low number of individuals that compose it and exchange of breeding stocks. The high Nei’s genetic distance between new goat populations with other breed indicative as a separate population while very low F_ST value among themselves indicate that they are significantly not differentiated with each other. The phylogeny and the plot for PCA indicated that Indian goat breeds can be grouped according to their physio-geographic location and production type.