Thumbnail Image

Thesis

Browse

 Reset filters

Settings

Subject: Animal Genetics and Breeding ×

Search Results

Now showing 1 - 9 of 19
  • Thumbnail Image
    ThesisItemOpen Access
    Exploration of Artificial Intelligence for the Prediction of Genetic Merit in Sheep
    (SKUAST Kashmir, 2022) Hamadani, Ambreen; Ganai, Nazir Ahmad
    As the amount of data on farms grows, it is important to evaluate the potential of artificial intelligence for making farming predictions. Considering all this, this study was undertaken to estimate genetic parameters as well as to evaluate various machine learning (ML) algorithms using 52-year data for sheep. Data preparation was done before analysis. Principal component regression (PCA) and feature selection (FS) were done and based on that, three datasets were created viz. PCA, PCA+ FS, and FS bodyweight prediction. Least-squares analysis and the effects of various factors on heritability were studied. Heritability was estimated using animal models with varying fixed and random effects. Breeding values were estimated using Best Linear Unbiased Prediction and accuracies based on varying fixed effects were compared. 14 ML algorithms were evaluated. All extreme outliers were removed through winsorization. The variance inflation factor for all features selected through PCA was 1. For the least-squares estimates, the effect of the sex was significant, parity insignificant and the effect of year of birth significant for most years. Both the additive genetic variance and the heritability got reduced upon adding more random effects. The inclusion of the year effect as a random variable reduced the heritability significantly except for birth weight. The highest average accuracy was observed for models having sex or sex and parity as fixed effects.The correlations between true and predicted values for artificial neural networks, Bayesian ridge regression, classification and regression trees, genetic algorithms, gradient boosting algorithm, K nearest neighbours, MARS algorithm, polynomial regression, principal component regression, random forests, support vector machines, XGBoost algorithm were 0.852, 0.742, 0.869, 0.762, 0.915, 0.781, 0.746, 0.742, 0.746, 0.917, 0.777, 0.915 respectively for breeding value prediction and 0.984, 0.992, 0.984, 0.734, 0.991, 0.949, 0.993, 0.957, 0.982, 0.99, 0.991, 0.99 respectively for bodyweights From this study it is concluded that for fixed and random factors often have a significant effect on the estimation of genetic factors and for the prediction of body weights and breeding values respectively. The top five algorithms for breeding value prediction were random forests, gradient boosting algorithm, XGBoost algorithm, classification and regression trees and artificial neural networks while for the prediction of bodyweights, they were MARS, Bayesian ridge regression, Ridge regression, Support Vector Machines and Gradient boosting algorithm. A total of 14 and 13 machine learning models were developed for the prediction of bodyweights and breeding values in sheep in the present study. It may be said that machine learning techniques can perform predictions with reasonable accuracies and can thus be viable alternatives to conventional strategies. Key words: Artificial Intelligence, Breeding Value,Genetic Parameters, Machine Learning
  • Thumbnail Image
    ThesisItemOpen Access
    Exploration of Artificial Intelligence for the Prediction of Genetic Merit in Sheep
    (SKUAST Kashmir, 2022) Hamadani, Ambreen; Ganai, Nazir Ahmad
    As the amount of data on farms grows, it is important to evaluate the potential of artificial intelligence for making farming predictions. Considering all this, this study was undertaken to estimate genetic parameters as well as to evaluate various machine learning (ML) algorithms using 52-year data for sheep. Data preparation was done before analysis. Principal component regression (PCA) and feature selection (FS) were done and based on that, three datasets were created viz. PCA, PCA+ FS, and FS bodyweight prediction. Least-squares analysis and the effects of various factors on heritability were studied. Heritability was estimated using animal models with varying fixed and random effects. Breeding values were estimated using Best Linear Unbiased Prediction and accuracies based on varying fixed effects were compared. 14 ML algorithms were evaluated. All extreme outliers were removed through winsorization. The variance inflation factor for all features selected through PCA was 1. For the least-squares estimates, the effect of the sex was significant, parity insignificant and the effect of year of birth significant for most years. Both the additive genetic variance and the heritability got reduced upon adding more random effects. The inclusion of the year effect as a random variable reduced the heritability significantly except for birth weight. The highest average accuracy was observed for models having sex or sex and parity as fixed effects.The correlations between true and predicted values for artificial neural networks, Bayesian ridge regression, classification and regression trees, genetic algorithms, gradient boosting algorithm, K nearest neighbours, MARS algorithm, polynomial regression, principal component regression, random forests, support vector machines, XGBoost algorithm were 0.852, 0.742, 0.869, 0.762, 0.915, 0.781, 0.746, 0.742, 0.746, 0.917, 0.777, 0.915 respectively for breeding value prediction and 0.984, 0.992, 0.984, 0.734, 0.991, 0.949, 0.993, 0.957, 0.982, 0.99, 0.991, 0.99 respectively for bodyweights From this study it is concluded that for fixed and random factors often have a significant effect on the estimation of genetic factors and for the prediction of body weights and breeding values respectively. The top five algorithms for breeding value prediction were random forests, gradient boosting algorithm, XGBoost algorithm, classification and regression trees and artificial neural networks while for the prediction of bodyweights, they were MARS, Bayesian ridge regression, Ridge regression, Support Vector Machines and Gradient boosting algorithm. A total of 14 and 13 machine learning models were developed for the prediction of bodyweights and breeding values in sheep in the present study. It may be said that machine learning techniques can perform predictions with reasonable accuracies and can thus be viable alternatives to conventional strategies. As the amount of data on farms grows, it is important to evaluate the potential of artificial intelligence for making farming predictions. Considering all this, this study was undertaken to estimate genetic parameters as well as to evaluate various machine learning (ML) algorithms using 52-year data for sheep. Data preparation was done before analysis. Principal component regression (PCA) and feature selection (FS) were done and based on that, three datasets were created viz. PCA, PCA+ FS, and FS bodyweight prediction. Least-squares analysis and the effects of various factors on heritability were studied. Heritability was estimated using animal models with varying fixed and random effects. Breeding values were estimated using Best Linear Unbiased Prediction and accuracies based on varying fixed effects were compared. 14 ML algorithms were evaluated. All extreme outliers were removed through winsorization. The variance inflation factor for all features selected through PCA was 1. For the least-squares estimates, the effect of the sex was significant, parity insignificant and the effect of year of birth significant for most years. Both the additive genetic variance and the heritability got reduced upon adding more random effects. The inclusion of the year effect as a random variable reduced the heritability significantly except for birth weight. The highest average accuracy was observed for models having sex or sex and parity as fixed effects.The correlations between true and predicted values for artificial neural networks, Bayesian ridge regression, classification and regression trees, genetic algorithms, gradient boosting algorithm, K nearest neighbours, MARS algorithm, polynomial regression, principal component regression, random forests, support vector machines, XGBoost algorithm were 0.852, 0.742, 0.869, 0.762, 0.915, 0.781, 0.746, 0.742, 0.746, 0.917, 0.777, 0.915 respectively for breeding value prediction and 0.984, 0.992, 0.984, 0.734, 0.991, 0.949, 0.993, 0.957, 0.982, 0.99, 0.991, 0.99 respectively for bodyweights From this study it is concluded that for fixed and random factors often have a significant effect on the estimation of genetic factors and for the prediction of body weights and breeding values respectively. The top five algorithms for breeding value prediction were random forests, gradient boosting algorithm, XGBoost algorithm, classification and regression trees and artificial neural networks while for the prediction of bodyweights, they were MARS, Bayesian ridge regression, Ridge regression, Support Vector Machines and Gradient boosting algorithm. A total of 14 and 13 machine learning models were developed for the prediction of bodyweights and breeding values in sheep in the present study. It may be said that machine learning techniques can perform predictions with reasonable accuracies and can thus be viable alternatives to conventional strategies.
  • Thumbnail Image
    ThesisItemOpen Access
    Development of MIS for Management and Evaluation of Sheep Breeding Data across Farms
    (SKUAST Kashmir, 2018) Hamadani, Ambreen; Ganai, Nazir Ahmad
    The science of Animal Breeding is based on the use of statistical designs and genetic models. Among other techniques, BLUP (Best Linear Unbiased Prediction) for breeding value (BV) estimation has a large impact in selection of animals. A fundamental requirement however, is the availability of accurate and reliable pedigreed data and computer aided tools for sophisticated computations. This necessitates the integration of Computer Aided Innovations into our current farming systems. Keeping this in view, Smart Sheep Breeder - A fully functional multi-use online Farm Management Information System (FMIS) for performance recording, farm data management and decision making in sheep farms was developed. The web-based tool facilitates performance recording in animals in simple to use data entry e-forms and generates customized reports on various aspects of sheep production. This web-based tool also estimates breeding values using Best Linear Unbiased Prediction (BLUP), calculates inbreeding coefficients, constructs Selection indices and generates pedigree and history sheets. In addition to the creation of the FMIS, Farm data from 4 Sheep Breeding Farms, viz. Mountain Research Station for Sheep and Goat, Shuhama and Government Sheep Breeding Farms, Kralpathri, Goabal and Reasi was analysed from 1969-2016, 1997-2016 and 1998 to 2008 respectively for various weight and wool traits. BLUP breeding values, phenotypic averages, heritability and inbreeding coefficients were estimated using the FMIS. Genetic, phenotypic, heritability trends were generated by linear regression on the year of birth. Inbreeding trends were also studied for Kashmir Merino flock. Most of the trends obtained for phenotypic values showed negative inclination. These trends were highly significant in Corriedale breed. Inbreeding coefficients showed an increasing trend while the breeding values and heritability were found to remain more or less static over the years. The trends indicate that very little genetic improvement has taken place over decades but the negative impact of management can be inferred from the declining phenotypic trends especially for Corriedale breed. The results suggested there is need for incorporation of smart tools for selection and breeding. Smart Sheep Breeder could thus prove to be indispensable for the present faming systems.
  • Thumbnail Image
    ThesisItemOpen Access
    Genetic Polymorphism of Foot Rot Gene Marker DQA2 Gene in Kashmir Merino and Corriedale Sheep
    (SKUAST Kashmir, 2020) Shayista, Akhter; Tavsief, Ahmad
    Foot rot is one of the most economically important contagious hoof diseases of sheep caused by mixed action of gram negative anerobes Dichelobacter nodusus and Fusobacterium necrophorum. Globally the disease results in production losses, welfare issues and negative economic implications with significant prevalence in Kashmir valley. Gene Ovine Leucocyte Antigen-DQA2 (OLA-DQA2) is one of the important genetic marker that has been found to be associated with foot-rot. The DQA2 gene is a member of Major Histocompatibility Complex (MHC) class-II, which is central to the innate immune response and mediates specific immunological responses to exogenous antigens. Keeping in view the importance of OLA-DQA2 gene, the present study was undertaken with the following objectives: a) To study different polymorphic variants of exon 2 of Ovar MHC class II DQA2 gene in Kashmir Merino and Corriedale Sheep. b) To characterise the allelic variants of this gene fragment for detection of single nucleotide polymorphism. To fulfill the above mentioned objectives single strand conformational polymorphism (SSCP) and DNA sequencing methods were used to detect polymorphism of DQA2 gene fragment. Furthermore experimental animal of interest, Kashmir Merino (n=30) maintained at Government Sheep Breeding Farm at Goabal Kangan, J&K and Corriedale sheep (n=30) maintained at Mountain Research Centre for Sheep and Goat (MRCS&G), F.V.Sc and A.H, Shuhama, SKUAST-K were included in the present study. Genomic DNA was isolated from blood samples by phenol-chloroform method and DNA quality and quantity was evaluated by agarose gel electrophoresis and UV spectrophotometer respectively. Exon 2 of DQA2 was amplified using nested polymerase chain reaction (PCR) by utizing degenerate primers based on previously gene published literature. Fragment of size 828 bp (comprising entire exon 2 flanked by 209 bp of intron 1 and 370 bp of intron 2) utilizing the primer set I and then product of size 242 bp fragment of DQA2 gene was successfully amplified utilizing the primer set II. PCR-SSCP technique for 242 bp fragment of exon 2 of DQA2 gene was carried out by 12% Polyacrylamide Gel Electrophoresis (PAGE). All the experimental animals were screened for SSCP and any distinctive pattern for a given amplicon was designated and scored as different genotype. Five different band patterns were observed in Kashmir Merino with genotype frequency of 0.10, 0.24, 0.13, 0.20 and 0.33 while, only three polymorphic variants were observed in corriedale sheep with genotype frequency of 0.33, 0.47 and 0.20. PCR products showing distinct banding pattern in PAGE gels were subjected to DNA sequencing to detect polymorphism at nucleotide level. The sequence data was subjected to bioinformatical analysis (Clustal W, Bio-Edit V7.0.5, Chromas V 2.6.5 and MEGA 5) to find any existed nucleotide variation and generation of phylogenetic trees. Sequence analysis revealed five alleles in Kashmir Merino and three alleles in Corriedale sheep in exon 2 of DQA2 gene. Comparision of allelic sequences and deduced amino acid translations suggests highly polymorphic nature of DQA2 gene due to various Single Nucleotide Polymorphism (SNP) which consequently results in change in amino acid composition and sequence. The results of the present study showed that exon 2 of DQA2 gene has rich genetic diversity and may be considered as an important marker for marker assisted selection in breeding of animals for disease resistance. The existence of polymorphism may be utilized in future to test the association with economic traits and sustainable selection program. Furthermore there is need to determine whether or not the currently commercially available gene test is indicative of foot rot resistance by taking association study of foot rot scoring system with molecular genetic markers in large sample size.
  • Thumbnail Image
    ThesisItemOpen Access
    Genetic Evaluation of Kashmir Merino Sheep at Organized Farms
    (SKUAST Kashmir, 2019) Rather, Mubashir Ali; Syed, Shanaz
    Kashmir Merino sheep is the major synthetic breed developed in the state of Jammu and Kashmir through crossing Gaddi, Bhakarwal and Poonchi with Tasmanian and Delain Merino at Reasi Jammu between 1942-1964 for apparel wool, better growth and adaptability. The present investigation was undertaken to evaluate the Kashmir Merino sheep for nine performance traits viz: birth weight (BWT), 6 months weight (6-MWT), yearling body weight (12-MWT), greasy fleece yield of first clip (GFY-1), greasy fleece yield of second clip (GFY-2), fiber diameter (FD), staple length (SL), age at first lambing (AFL) and Inter-lambing period (ILP) so as have an insight into performance of the breed over years. The data on 6300 birth records spread over twenty one years of Kashmir Merino sheep maintained at Government Sheep Breeding Farm Kralapathri and Government Sheep Breeding Farm Goabal Kashmir was analyzed with the Mixed Model Least Squares and Maximum Likelihood algorithms, PC-2 version computer programme (Harvey. 1990) to assess the random effect of sire and fixed effects of period, sex, farm, type of birth and parity on various production and reproduction traits. The overall least-squares means for BWT, 6-MWT, 12-MWT, GFY-1, GFY-2, FD, SL, AFL and ILP were 3.34±0.05 kg, 19.33±0.45 kg, 22.44±0.46 kg, 0.82±0.03 kg, 0.80±0.02 kg, 20.33±0.05 µ, 3.86±0.14 cms, 1090.22±19.45 days and 401.45±22.29 days, respectively. The sire and period of birth had highly (p<0.01) significant effect on all the performance traits except ILP. However, all the traits presented fluctuating trends over the periods. Effect of gender was highly significant (p<0.01) on all growth and wool quality traits. The male lambs were significantly heavier than females at all ages. The effect of parity of dam was highly significant (p<0.01) on the growth traits. The effect of birth type was non-significant on all traits under study. The effect of farm was significant (p<0.05) on the wool yield traits. The heritability estimates for BWT, 6-MWT, and 12-MWT, GFY-1, GFY-2, FD, SL, AFL and ILP were 0.49±0.07, 0.57±0.07, 0.58±0.07, 0.61±0.08, 0.52±0.04, 0.23±0.04, 0.66±0.10, 0.06±0.06 and 0.15±0.09 respectively. The genetic correlation ranged from -0.73±0.10 to 0.87±0.04 while as phenotypic correlation ranged from -0.66±0.07 to 0.54±0.01. The genetic trends were negative for BWT, 6-MWT, GFY-2, SL & AFL and positive for12-MWT, GFY-2, FD and ILP. The phenotypic trends were negative for BWT, GFY-2, FD, SL and AFL and 6-MWT, 12-MWT, GFY-1 and ILP. Average estimates of body growth and greasy fleece yield traits indicated a decline in performance. The overall least squares mean of 20.33±0.05 µ for fiber diameter indicates that Kashmir Merino sheep is comparable to some of the finest wool breeds of the world. Keeping in view the heritability and genetic correlations among performance traits it is concluded that 6-MWT can serve as a good selection criterion in this breed at early age. For making sound policy decisions the availability of accurate real-time date is essential for futuristic policy and planning. It is therefore essential to implement advanced ICT and breeding tools to the present farming system. This shell also ensures accuracy while significantly reducing task time, improving the accuracy and reliability of the data analysis.
  • Thumbnail Image
    ThesisItemOpen Access
    Genetic Evaluation and Trends of Growth and Age at First Lambing in a Correidale Flock of Sheep in Kashmir
    (SKUAST Kashmir, 2016) Baba, Mohd Ashraf; Ganai, T.A.S
    The present study was undertaken to examine the factors affecting some performance traits, estimate genetic and phenotypic parameters of these traits and trends in a Corriedale flock of sheep maintained at Sheep Research Station Shuhama, (SKUAST-K). A total of 14214 records of animals sired by 77 rams and spanning over a period of 25 years (1989–2013), were collected and analyzed. The analyzed traits included body weight at birth (BW), weaning (WW), six months (W6), twelve months (W12), eighteen months (W18) and age at first lambing (AFL). Owing to uneqal and non orthogonal nature of data, least squares technique was applied to assess the effect of non-genetic factors on body weights and AFL whereas genetic parameters and breeding values were estimated by using the DFREML software programme. The estimated overall least square means for BW, WW, W6, W12, W18 (kg) and age at first lambing (days) were 3.20 ± 0.091, 11.05 ±0.408, 16.06±0.452, 21.69±0.722, 33.425±0.776 and 882.50±11.33, respectively.Year of birth had a highly significant (P<0.01) influence on all the studied traits. The effect of type of the birth was significant on BW,WW (P<0.01),W6 (P<0.05) with no effect on W12,W18 and AFL. Sex of the lamb had a significant (P<0.01) effect on BW, WW, W6 and W12 and no significant effect on W18. Dam age had a significant effect (P<0.01) on all the weight traits except W18. It also had a significant (P<0.05) effect on AFL. Direct heritability estimates for BW, WW and W6 were 0.11± 0.03, 0.19±0.04, 0.29± 0.05 whereas maternal heritability estimates for the same traits were 0.24±0.04, 0.12±0.03 and 0.07±0.04 respectively. Direct heritability estimates for W12, W18 and AFL were 0.38±0.07, 0.40± 0.06 and 0.11±0.06 respectively. Genetic and phenotypic correlations among all growth traits were positive and were low to moderately high in magnitude while the correlations between any of the growth trait and AFL were negatively. The correlations between any adjacent growth traits were higher than non adjacent traits. A statistically significant (P<0.01) positive genetic trend was observed for BW,WW,W6 andW12 body weights while as for W18 and AFL it was nonsignificant (P<0.05). Phenotypic trends for all the traits under study were statistically nonsignificant (P<0.05). The overview of the study reveals that genetic and non genetic factors affect the growth and fertility and they should be taken into consideration while evaluating the animals for performance traits. Positive and moderate to high genetic correlations among most of the body weight traits indicates the possibility of correlated response through multiple trait selection. The magnitude of heritability for the traits especially for relatively mature body weights indicate that there is further scope of improvement through selective breeding with emphasis on these traits. Irregular genetic and phenotypic trends depicted for the examined years might reveal the importance of implementing selection based on breeding values and there is further scope of improvement if reasonable levels of management and selection pressure are maintained.
  • Thumbnail Image
    ThesisItemOpen Access
    Characterization of Exon VIII of BMPR-1B Gene and Effects of Mutation in the region on Growth & Reproductive Traits of Sheep with BMPR-1B Gene in Kashmir Sheep
    (SKUAST Kashmir, 2015) Sheikh, Amir Abdullah; Ganai, T.A.S
    High prolificacy in sheep has been observed among different breeds and even within the same breed. Genetic studies have established that in some cases ovulation rates and litter size can be genetically determined by the action of single gene(s) called Fec genes with a major effect. BMPR1B gene is located on chromosome 6 and consists of 12 exons. Single nucleotide polymormpism (SNP) has been identified at nucleotide position of 746 of cDNA. The present study was undertaken to perform Characterization of Exon VIII of BMPR 1B Gene and Its effects on Growth & Reproductive Traits of Sheep with FecB gene Polymerase Chain Reaction-Restriction Fragment Length Polymorphism (PCR-RFLP) and Polymerase Chain Reaction-Single Strand Conformation Polymorphism (PCR-SSCP) techniques were used in this study. The investigations in present study pertain to 140 bp fragment of Exon-VIII BMP1B gene. Two SSCP genotypes (B+ and ++) of two designated alleles (B and +) were identified for exon-8. Sequence analysis of different SSCP variants revealed mutations present in them. Two groups were selected-one in which FecB gene was introgressed (Test group) and the other group was wild (control group). Mutation in designated allele B was observed to be specific to the Kashmir valley sheep introgressed with FecB gene (Test group) and was not observed in control group (without FecB introgression). The overall frequencies of B+ and ++ genotypes of exon-8, were 0.38 and 0.62, respectively for test group. In case of control group, the overall frequencies of B+ and ++ genotypes, were 0 and 1, respectively. The results indicated that crossing of Kashmir breed ewes with FecB rams leads to a reduction in the body weight of crossbred lambs than the pure Kashmir sheep, but the overall productivity of crossbred was improved over Kashmir sheep. The effect of polymorphism of exon-8 was significant (P0.05). The overall mean of AFL for Fec B introgressed sheep was 705.15 ±19.00 days while as for control group it was 769.23 ±14.40 days.
  • Thumbnail Image
    ThesisItemOpen Access
    Molecular Characterization of Exon-VII and Screening of Breeding Population for Allele A and Allele B of β-Casein Gene of Cattle
    (SKUAST Kashmir, 2015) Najar, Mudasir Ahmad; Ganai, T.A.S
    The present study was undertaken to perform characterization of exon-VII and screening of breeding population of Holstein friesen bulls, Jersey bulls, crossbred cows and local cattle of Jammu & Kashmir state for allele A1 & A2 of beta casein gene. A wide variation in the beta casein allele A1 & A2 frequency among the Holstein friesen, Jersey, crossbred and local cattle of Jammu & Kashmir was found suggesting that the molecular selection for animals carrying the allele A1 could impact breeding programs for dairy production. The genotyping of CSN2 alleles A1 and A2 is of the practical importance since the CSN2A1 is associated with the liberation of β-casomorphin-7 and other bioactive peptides with opioid nature and thus cause of several human non-communicable diseases. The A1A1 genotype was most prevalent in Holstein friesen bulls (0.63), followed by Jersey bulls (0.15) and crossbred cows (0.13), whereas it was observed to be lowest in local cattle of Jammu & Kashmir. The heterozygote A1A2 genotype was most abundant in crossbred cows (0.75), followed by Jersey bulls (0.31), HF bulls (0.22) and local cattle (0.20). The A2A2 genotype was most frequent in local cattle (0.8), followed Jersey bulls (0.54), HF bulls (0.15) and crossbred cows (0.12). The overall frequency of three designated genotypes A1A1, A1A2 and A2A2 across the four genetic groups of cattle were 0.25, 0.52 and 0.23 respectively. The frequency of A1 allele was the highest in HF bulls (0.74), followed by crossbred cows (0.51), Jersey bulls (0.31) and relatively low in local cattle (0.1). The local cattle has the highest (0.9) frequency of A2 allele, where as its frequencies were 0.49, 0.69 and 0.26 in crossbred cows, Jersey and HF bulls respectively. The overall frequencies of two alleles A1 and A2 in whole population studied were 0.51 and 0.49 respectively. The relatively high incidence of the CSN2 allele A1 is characteristic for the breeding populations and this special allele distribution could be used to develop selection strategies to breed specialized lines of exotic breeds with local cows. The presence of A2 allele in high frequency in farmer’s herd indicates that this allele being without any adverse effect on human health needs to be conserved in the crossbred and graded animals. Similarly, increasing frequency of A1 allele in crossbred cows was mainly due to indiscriminate crossbreeding with exotic breeds and thus when native breeds are crossed or upgraded by economically important Bos taurus breeds, the breeding bull must be screened for A1 allele of β-casein gene and if a breeding bull possesses this allele it should be taken out of breeding programmes.
  • Thumbnail Image
    ThesisItemOpen Access
    Genetic Diversity among Cashmere (Pashmina) Goat population of India through Microsatellite Markers Approach
    (SKUAST Kashmir, 2014) Bhat, Mushtaq Ahmad; Ganai, T.A.S
    The present study was undertaken for population genetic analysis of microsatellite variation in three Cashmere (Pashmina) producing goat populations of India viz: Chegu, Tibetain and Changthangi using 10 microsatellite markers selected from the list suggested by ISAG-FAO. Observed number of alleles ranged from three (in ILSTS-005and-065) to seven (in ILSTS-008) with total 54 alleles across three populations. The overall heterozygosity, PIC and Shannon index values were 0.57, 0.64 and 1.49, respectively, indicating high gene diversity. The highest observed heterozygosity was found in Chegu and minimum in Tibetain goat. Genetic distance was least (0.2045) between Changthangi and Tibetain and the highest between Chegu and Changthangi (0.3621). In all populations moderate inbreeding was indicated (mean FIS = 0.1664, FIT = 0.2280) within and among the breeds. Genetic differentiation between breeds was moderate with a mean FST value of 0.0738, which showed that an average proportion of genetic variation explained by breed differences was 7.3%, the remaining 92.62% corresponding to differences among individuals. The clusters obtained on phylogenetic tree generated from Nei’s genetic distance matrix agreed with the geographic origin of the breed. The genetic relatedness between Changthangi and Tibetain breeds indicates intermixing of these breeds or same origin of these breeds. Deviations from Hardy-Weinberg equilibrium were noted for most of the loci. Bottleneck analysis revealed bottleneck for Chegu population under all the three mutation models for sign test, whereas the standardised difference test and Wilcoxon rank test indicated heterozygosity excess in Chegu population under TPM and SMM models, but bottleneck was observed under IAM mutation model. Tibetain exhibited heterozygosity excess under all the three mutation models for sign test, standardised difference test and Wilcoxon rank test. Similar results were observed for Changthangi with all the three tests under IAM, TPM and SMM models. Hence results from sign test, standardised difference test and Wilcoxon rank test revealed bottleneck in Chegu goat, where as Tibetain and Changthangi goat were free from any bottleneck. These 10 markers can be used for genetic investigations and assessing population structure in Indian goat populations. The study analyzed the population structure of these populations and contributed to the knowledge and genetic characterization of three Cashmere (Pashmina) goat populations. In addition, the microsatellites recommended by ISAG proved to be useful for the biodiversity studies in goat germplasm.