Gene expression: Nguni and Bonsmara cattle

A gene-expression study on the growth performance of Nguni and Bonsmara cattle grown in a feedlot fed a high and low energy diet.

Industry Sector: Cattle And Small Stock

Research Focus Area: Livestock production with global competitiveness: Breeding, physiology and management

Research Institute: Agricultural Research Council

Year Of Completion : 2019

Researcher: Dina Linde

The Research Team

TitleInitialsSurnameHighest QualificationResearch Institution
DrM.M.ScholtzPhDARC-AP
ProfEvan Marle-KosterPhDUP
MrsA. Theunissen MSc Vaalharts Research Station

Executive Summary

Introduction

In South Africa, Nguni cattle are one of the breeds found predominantly in extensive production systems. The Nguni is an indigenous cattle breed and is widely used in crossbreeding systems due to their high fertility and mothering ability. Nguni cattle are also commonly used in communal production systems. The majority of beef in South Africa is produced in feedlots using commercially formulated high energy diets, where preference is given to medium and large framed later maturing cattle that include British types and composites such as the South African Bonsmara. Due to the Nguni’s small frame and low meat yield when compared to British types it is nor preferred as a feedlot animals, however studies have shown that Nguni cattle produce high quality meat. The veldt of South Africa has a varying degree of carrying capacity, however the occasional drought conditions have necessitated the use of alternative production systems such as feedlots for finishing cattle. The use of a lower energy diet in feedlots for indigenous cattle have been suggested and warrants investigation. Nutrigenomics is the study of the effect of nutrition on the genes of the animal by quantifying the gene expression. An improved understanding of the interaction between the nutritional environment and the genetics of the animal can lead to increased efficiency and production. Diets that are different in components or ingredients can result in different phenotypes in the animals. In this study the effect of two feedlot diets with different energy levels have been investigated using a transcriptome approach.

Objective statement

The objective of this study is to determine if there are gene expression differences between Nguni and Bonsmara cattle fed a low or a high energy diet. The gene expression of the cattle can determine the underlying differences in the reaction of the two breeds to the two different diets.

Project Aims

  1. To evaluate whether the Celtic mutation on the POLL locus is the causative mutation for polledness in Bonsmara and Drakensberger
  2. To perform a genome wide association study of the Polled and Scur genes based on phenotypic data and genotypic data from the GGP Bovine 150K SNP bead chip
  3. To apply sequence data available from the Bovine Genomics Program to finemap the suspected regions for the Polled and Scur genes

Results

Performance results showed a higher live weight, carcass weight and marbling score for all bulls fed the high energy diet compared to bulls fed the low energy diet. Only carcass weight and marbling score had significant difference in terms of diet (p < 0.05). Live weight, average daily gain, rib fat, rump fat and eye muscle area were only significant for breed. Diet had a greater effect on the Bonsmara compared to the Nguni according to transcriptomic and phenotypic values. Transcriptomic values showed 3584 differential expressed genes (DEG) between the Bonsmara fed the two different diets, while only a difference of 288 DEG were observed between the Nguni fed the two different diets. Phenotypic values show a difference of 20 kg between the Bonsmara groups and only a 6 kg difference between the Nguni groups. Most DEG were involved with cellular processes and metabolic pathways. A total of 73 differentially expressed genes were observed between the diets across breeds. The genes that were involved in intramuscular fat deposition (CRHR2, NR4A3, MMD) were expressed on a higher level in the bulls on the low energy diet compared to bulls on the high energy diet. Genes that were involved in muscle deposition (PITX2, Leptin, AVP) was expressed higher in the bulls on the high energy diet. Comparing the breeds revealed that 2214 genes were differentially expressed between the Bonsmara and the Nguni. At the end of the feedlot trial a higher expression of marbling genes (SIRT, ND, ADIPOQ) were observed in the Nguni, however this expression was not observed in the marbling scores recorded. Several genes (ASIP, MOGAT, SNAI3) that were involved in fat deposition were upregulated in the Bonsmara. This suggests that the Nguni was still growing at the end of the feedlot trial while the Bonsmara had reached physiological maturity. Little literature could be found on some of the gene showing the highest expression in the groups such as GSTA3, TEX28 and TUBB3. Glutathione-s transferase alpha 3 (GSTA3) is linked to steroidal genesis and could therefore have an influence on myogenesis, however no confirming literature could be found.

Conclusion

The diet had a smaller effect on the Nguni bulls compared to the Bonsmara bulls as observed in the DEG and the carcass weight of the bulls. This might indicate that the Nguni is more adaptable to a variation in feed quality. The Bonsmara bulls had a higher meat yield on the high energy diet, however it seems as if the bulls on the low energy diet had a higher expression of the marbling genes. This is in contrast to the phenotypically higher marbling score. Further research needs to be done as this study had a small sample size (n=40). An extended feedlot period for Nguni cattle should be considered in future studies. This study provides reference data for differentially expressed genes in muscle of South African feedlot cattle.

Popular Article

A low energy feedlot diet may favour our indigenous breeds

Dina Linde1,2, Michiel Scholtz1,3 & Este van Marle-Koster2

Introduction

The phenotypes of animals vary due to differences in their genetics and environment. The expression of the genes of animals can also be influenced by their diet. In South Africa, Sanga cattle (Afrikaner and Nguni) are adapted to various environments and production systems. However, these cattle are mostly found in extensive production systems that make use of natural grazing. They are not preferred in the feedlot, as their smaller frame sizes compared to Bos Taurus types and crossbreds will lead to smaller carcass sizes. The traditional diet fed in the feedlot is a high energy diet, that includes maize and maize by-products. There are however, farmers that believe that a high energy diet does not suite the genetics of Sanga cattle and that these breeds would do better on a lower energy diet.

A study was done to find alternative strategies for Sanga cattle in the feedlot. A diet low in energy (10.9 MJ ME/kg) or a high energy diet (12.5 MJ ME/kg) was fed to Nguni and Bonsmara cattle for a period of a 120 days respectively. During the feeding period the animals were weighed and real time ultrasound scanned for meat quality characteristics such as eye muscle area (EMA). At slaughter, muscle samples were taken to analyse the gene expression data.

In terms of the feedlot data, no significant difference could be found between the diets with regard to live weight, eye muscle area (EMA), rib fat and rump fat in the Nguni. Breed differences were however found. Carcass weight and marbling did however show significant difference as well as an interaction between breed and diet.

Breed had a much bigger effect compared to diet, with 2214 differentially expressed genes (DEG) and 74 DEG, respectively. Of the  genes found differentially expressed between the breeds, several genes known to be involved in marbling (SIRT, ND, COX, ADIPOQ, SERPINF2) was expressed higher in the Nguni compared to the Bonsmara. However, the phenotypic marbling score was higher in the Bonsmara compared to the Nguni.

This lead to the suggestion that the Nguni needed more time (a longer feedlot period) for the expression of the marbling genes to show phenotypically. This is in contrast to the industry that perceives that the Nguni deposits fat too early in comparison to exotic and crossbreds. Sanga cattle tend to first deposit fat intramuscularly, which might be an adaptation mechanism for the harsh conditions in which these cattle lived. It seems, however, that the Nguni only begins depositing muscle tissue after a sufficient layer of fat is deposited. This might be the reason for backgrounding, which is common practise in the Nguni breed.

Between the Nguni fed the high energy diet and the Nguni fed the low energy diet, 288 genes were differentially expressed. The different levels of energy in the diets seem to result in different components being deposited. Various genes (PITX2, PAX, Leptin, AVP, OXT) involved in muscle deposition were upregulated in the bulls that received the high energy diet, compared to the bulls fed the low energy diet. This is also seen in the phenotypic results with the carcass weight of the bulls fed the high energy diet being higher compared to the bulls fed the low energy diet.

The difference in carcass weight is very small between the Nguni fed the high energy diet and the Nguni fed the lower energy diet (6 kg). However, genes that influence intramuscular fat deposition (SPARC, CRNR2, CHRND, NR4A3, MMD) were elevated in the bulls that received the low energy diet compared to the bulls that received the high energy diet. This was also not shown in the phenotypic traits. Extending the period of feeding may result in the bulls that received the low energy diet to express the phenotype.

It seems that the low energy diet suite the Sanga cattle better, when compared to the traditional high energy diet fed in the feedlots. This should be further investigated as this study had a relatively small sample size (n=40). Furthermore, as these animals are ruminants, it would be also interesting to study the rumen microbiome in another study of this kind.

Dina Linde1,2, Michiel Scholtz1,3 & Este van Marle-Koster2

1ARC – Animal Production, Private bag X2, Irene, 0062, South Africa; 2Department of Animal and Wildlife science, University of Pretoria, Pretoria, 0002, South Africa; 3Department of Animal, Wildlife and Grassland Sciences, University of the Free State, Bloemfontein, 9300; South Africa;

Footer and Tags and Categorisation

Please contact the Primary Researcher if you need a copy of the comprehensive report of this project on :LindeD@arc.agric.za

Gene expression: Nguni and Bonsmara

A gene-expression study on the growth performance of Nguni and Bonsmara cattle grown in a feedlot fed a high and low energy diet

Industry Sector: Cattle And Small Stock

Research Focus Area: Livestock production with global competitiveness: Breeding,physiology and management

Research Institute: Agricultural Research Council

Year Of Completion : 2019

Researcher: Dina Linde

The Research Team

TitleInitialsSurnameHighest QualificationResearch Institution
DrM.M.ScholtzPhDARC-AP
ProfEvan Marle-KosterPhDUP
MrsA. Theunissen MSc Vaalharts Research Station

Executive Summary

Introduction

In South Africa, Nguni cattle are one of the breeds found predominantly in extensive production systems. The Nguni is an indigenous cattle breed and is widely used in crossbreeding systems due to their high fertility and mothering ability. Nguni cattle are also commonly used in communal production systems. The majority of beef in South Africa is produced in feedlots using commercially formulated high energy diets, where preference is given to medium and large framed later maturing cattle that include British types and composites such as the South African Bonsmara. Due to the Nguni’s small frame and low meat yield when compared to British types it is nor preferred as a feedlot animals, however studies have shown that Nguni cattle produce high quality meat.

The veldt of South Africa has a varying degree of carrying capacity, however the occasional drought conditions have necessitated the use of alternative production systems such as feedlots for finishing cattle. The use of a lower energy diet in feedlots for indigenous cattle have been suggested and warrants investigation. Nutrigenomics is the study of the effect of nutrition on the genes of the animal by quantifying the gene expression. An improved understanding of the interaction between the nutritional environment and the genetics of the animal can lead to increased efficiency and production. Diets that are different in components or ingredients can result in different phenotypes in the animals. In this study the effect of two feedlot diets with different energy levels have been investigated using a transcriptome approach.

Objective statement

The objective of this study is to determine if there are gene expression differences between Nguni and Bonsmara cattle fed a low or a high energy diet. The gene expression of the cattle can determine the underlying differences in the reaction of the two breeds to the two different diets.

Project Aims

  1. To evaluate whether the Celtic mutation on the POLL locus is the causative mutation for polledness in Bonsmara and Drakensberger
  2. To perform a genome wide association study of the Polled and Scur genes based on phenotypic data and genotypic data from the GGP Bovine 150K SNP bead chip
  3. To apply sequence data available from the Bovine Genomics Program to finemap the suspected regions for the Polled and Scur genes

Results

Conclusion

Performance results showed a higher live weight, carcass weight and marbling score for all bulls fed the high energy diet compared to bulls fed the low energy diet. Only carcass weight and marbling score had significant difference in terms of diet (p < 0.05). Live weight, average daily gain, rib fat, rump fat and eye muscle area were only significant for breed. Diet had a greater effect on the Bonsmara compared to the Nguni according to transcriptomic and phenotypic values. Transcriptomic values showed 3584 differential expressed genes (DEG) between the Bonsmara fed the two different diets, while only a difference of 288 DEG were observed between the Nguni fed the two different diets. Phenotypic values show a difference of 20 kg between the Bonsmara groups and only a 6 kg difference between the Nguni groups. Most DEG were involved with cellular processes and metabolic pathways. A total of 73 differentially expressed genes were observed between the diets across breeds. The genes that were involved in intramuscular fat deposition (CRHR2, NR4A3, MMD) were expressed on a higher level in the bulls on the low energy diet compared to bulls on the high energy diet. Genes that were involved in muscle deposition (PITX2, Leptin, AVP) was expressed higher in the bulls on the high energy diet. Comparing the breeds revealed that 2214 genes were differentially expressed between the Bonsmara and the Nguni.

At the end of the feedlot trial a higher expression of marbling genes (SIRT, ND, ADIPOQ) were observed in the Nguni, however this expression was not observed in the marbling scores recorded. Several genes (ASIP, MOGAT, SNAI3) that were involved in fat deposition were upregulated in the Bonsmara. This suggests that the Nguni was still growing at the end of the feedlot trial while the Bonsmara had reached physiological maturity. Little literature could be found on some of the gene showing the highest expression in the groups such as GSTA3, TEX28 and TUBB3. Glutathione-s transferase alpha 3 (GSTA3) is linked to steroidal genesis and could therefore have an influence on myogenesis, however no confirming literature could be found.

Conclusion

The diet had a smaller effect on the Nguni bulls compared to the Bonsmara bulls as observed in the DEG and the carcass weight of the bulls. This might indicate that the Nguni is more adaptable to a variation in feed quality. The Bonsmara bulls had a higher meat yield on the high energy diet, however it seems as if the bulls on the low energy diet had a higher expression of the marbling genes. This is in contrast to the phenotypically higher marbling score. Further research needs to be done as this study had a small sample size (n=40). An extended feedlot period for Nguni cattle should be considered in future studies. This study provides reference data for differentially expressed genes in muscle of South African feedlot cattle.

Popular Article – A low energy feedlot diet may favour our indigenous breeds

Dina Linde1,2, Michiel Scholtz1,3 & Este van Marle-Koster2

1ARC – Animal Production, Private Bag X2, Irene, 0062, South Africa; 2Department Of Animal And Wildlife Science, University Of Pretoria, Pretoria, 0002, South Africa; 3Department Of Animal, Wildlife And Grassland Sciences, University Of The Free State, Bloemfontein, 9300; South Africa;

The phenotypes of animals vary due to differences in their genetics and environment. The expression of the genes of animals can also be influenced by their diet. In South Africa, Sanga cattle (Afrikaner and Nguni) are adapted to various environments and production systems. However, these cattle are mostly found in extensive production systems that make use of natural grazing. They are not preferred in the feedlot, as their smaller frame sizes compared to Bos Taurus types and crossbreds will lead to smaller carcass sizes. The traditional diet fed in the feedlot is a high energy diet, that includes maize and maize by-products. There are however, farmers that believe that a high energy diet does not suite the genetics of Sanga cattle and that these breeds would do better on a lower energy diet.

A study was done to find alternative strategies for Sanga cattle in the feedlot. A diet low in energy (10.9 MJ ME/kg) or a high energy diet (12.5 MJ ME/kg) was fed to Nguni and Bonsmara cattle for a period of a 120 days respectively. During the feeding period the animals were weighed and real time ultrasound scanned for meat quality characteristics such as eye muscle area (EMA). At slaughter, muscle samples were taken to analyse the gene expression data.

In terms of the feedlot data, no significant difference could be found between the diets with regard to live weight, eye muscle area (EMA), rib fat and rump fat in the Nguni. Breed differences were however found. Carcass weight and marbling did however show significant difference as well as an interaction between breed and diet.

Breed had a much bigger effect compared to diet, with 2214 differentially expressed genes (DEG) and 74 DEG, respectively. Of the  genes found differentially expressed between the breeds, several genes known to be involved in marbling (SIRT, ND, COX, ADIPOQ, SERPINF2) was expressed higher in the Nguni compared to the Bonsmara. However, the phenotypic marbling score was higher in the Bonsmara compared to the Nguni.

This lead to the suggestion that the Nguni needed more time (a longer feedlot period) for the expression of the marbling genes to show phenotypically. This is in contrast to the industry that perceives that the Nguni deposits fat too early in comparison to exotic and crossbreds. Sanga cattle tend to first deposit fat intramuscularly, which might be an adaptation mechanism for the harsh conditions in which these cattle lived. It seems, however, that the Nguni only begins depositing muscle tissue after a sufficient layer of fat is deposited. This might be the reason for backgrounding, which is common practise in the Nguni breed.

Between the Nguni fed the high energy diet and the Nguni fed the low energy diet, 288 genes were differentially expressed. The different levels of energy in the diets seem to result in different components being deposited. Various genes (PITX2, PAX, Leptin, AVP, OXT) involved in muscle deposition were upregulated in the bulls that received the high energy diet, compared to the bulls fed the low energy diet. This is also seen in the phenotypic results with the carcass weight of the bulls fed the high energy diet being higher compared to the bulls fed the low energy diet.

The difference in carcass weight is very small between the Nguni fed the high energy diet and the Nguni fed the lower energy diet (6 kg). However, genes that influence intramuscular fat deposition (SPARC, CRNR2, CHRND, NR4A3, MMD) were elevated in the bulls that received the low energy diet compared to the bulls that received the high energy diet. This was also not shown in the phenotypic traits. Extending the period of feeding may result in the bulls that received the low energy diet to express the phenotype.

It seems that the low energy diet suite the Sanga cattle better, when compared to the traditional high energy diet fed in the feedlots. This should be further investigated as this study had a relatively small sample size (n=40). Furthermore, as these animals are ruminants, it would be also interesting to study the rumen microbiome in another study of this kind.

Footer and Tags and Categorisation

Please contact the Primary Researcher if you need a copy of the comprehensive report of this project on :LindeD@arc.agric.za

Genotype imputation for indigenous beef cattle

Genotype imputation as a genomic strategy for the South African Drakensberger beef breed

Industry Sector: Cattle And Small Stock

Research Focus Area: Livestock Production With Global Competitiveness: Breeding,Physiology And Management

Research Institute: Department Of Agriculture Forest And Fisheries (DAFF)

Year Of Completion : 2019

Researcher: Carina Visser

The Research Team

TitleInitialsSurnameHighest QualificationResearch Institution
DrM.M.ScholtzPhDARC-AP
ProfEvan Marle-KosterPhDUP
MrsA. Theunissen MSc Vaalharts Research Station

Executive Summary

The SA Drakensberger is a medium-framed breed with a sleek, black coat. Considering its history as one of the oldest indigenous breeds, its prominent role in the present beef industry and its potential for improving the beef cattle gene pool in the future; there is value in characterizing the SA Drakensberger on the genomic level. There has recently been interest in incorporating genomic information into selection strategies for this breed. Apart from the fact that the implementation of genomic technologies relies on diligent phenotyping efforts, accurate and complete pedigree recording; genomic selection also requires adequate SNP genotyping profiles (Meuwissen et al., 2001). The SA Drakensberger meets the requirements for genomic selection with 100% participation in SA Stud Book’s Logix Beef performance recording scheme as well as an extensive recorded pedigree profile (SA Stud Book, 2017). Theoretically, current EBVs can therefore be enhanced with the use of genomics if financial resources allow the generation of adequate high-density genotypic profiles. Imputation is a statistical methodology that relies on the genomic segments shared within a breed, or a group of genetically similar breeds, to predict genotypic information for SNPs that were not physically genotyped (Marchini et al., 2007). The main advantage of this methodology is the reduction in genotyping costs by allowing genotyping to be undertaken using lower density SNP panels. The utility of such low-density panels for applications such as genomic selection will depend on the accuracy with which un-genotyped SNPs can be imputed to higher density from such lower density panels. Even though imputation is integrated into routine genomic evaluations internationally, the utility of this methodology has not been evaluated for indigenous cattle resources. Considering that these breeds often have admixed genomes, applying imputation requires optimization for such breeds and this includes the SA Drakensberger.

Objective Statement

The objective of this research project was to comprehensively study the validity of genotype imputation, from lower-density single nucleotide polymorphism (SNP) panels to higher density, for the economically-important SA Drakensberger beef cattle breed towards cost-effectively implementing genomically-enhanced breed improvement strategies such as genomic selection for this indigenous breed in the future.

Project Aims

  1. To evaluate whether the Celtic mutation on the POLL locus is the causative mutation for polledness in Bonsmara and Drakensberger
  2. To perform a genome wide association study of the Polled and Scur genes based on phenotypic data and genotypic data from the GGP Bovine 150K SNP bead chip
  3. To apply sequence data available from the Bovine Genomics Program to finemap the suspected regions for the Polled and Scur genes

Results

Results generated from the first part of this study indicated that differences in genomic characteristics such as minor allele frequency (MAF), linkage disequilibrium (LD) and runs of homozygosity (ROH) exists between chromosomes. Mean genome-wide MAF was, for example, estimated to be 0.26 with chromosome-specific MAF ranging from 0.24 (Bos Taurus Autosome; BTA14) to 0.28 (BTA21). This was supported by the proportion of low-MAF (< 5%) SNPs estimated, which indicated 16.0% of SNPs to be classified as low-MAF SNPs on BTA14. The inter-SNP LD was generally weak, ranging from mean r²=0.11 (BTA28) to r²=0.17 (BTA14) for SNPs separated by≤1Mb and r²=0.20 extended only up to<30 kb. LD was weaker between SNP pairs including low-MAF SNPs. Consensus ROH segments were identified and the most prevalent of these occurred on BTA14 and was identified in ∼23% of the sampled population. The ROH length characteristics furthermore pointed towards more ancient inbreeding, reflecting known historic bottleneck events.

For the second and main object preliminary results were generated to understand the necessary dynamics, in terms of size and composition, of an appropriate sub-population to use as a reference for estimation of haplotypes to be imputed from. Initial results indicated that a larger reference population would improve imputation accuracy. For example, it was observed that a 4% increase in imputation accuracy could achieved when the ratio of reference:test population was 90:10 versus 75:25; imputation accuracy improved from 0.981 (range: 0.895-0.997) to 0.985 (range: 0.905-0.996) when the former versus the latter scenario was used. It was further observed that using a reference population consisting of animals with closer genetic relatedness to the test population would also improve imputation accuracy. A strong correlation of 0.817 (P<0.001) was observed between the mean genetic relatedness of animals in the test population, with animals in the reference population, and their resulting imputation accuracy.

This was supported by estimates showing mean imputation accuracy of 0.994 as opposed to 0.982 for animals that had both as opposed to no parents in the reference population. The influence of using different low-density SNP panels, consisting of varying density and SNP content, on more specifically animal-wise and SNP-wise imputation accuracy was then determined. Animal-wise imputation accuracy improved when the SNP density of the lower-density panel improved; correlation-based imputation accuracy ranged (minimum to maximum) from 0.625-0.990, 0.728-0.994, 0.830-0.996, 0.885-0.998 and 0.918-0.999 when 2 500, 5 000, 10 000, 20 000 and 50 000 SNPs when SNPs were randomly chosen. The variation between animals, as well as the degree of improvement in accuracy, became smaller with increasing SNP density. Improvements of 0.043 units were seen when SNPs were doubled from 2 500 to 5 000 SNPs, as opposed to an improvement 0f only 0.007 units when SNPs were (more than) doubled from 20 000 to 50 000 SNPs. Selection of SNPs based on both MAF and LD attributes proved to be the best selection strategy to maximize imputation accuracy and random selection produced the worst imputation accuracy. Mean imputation accuracy exceeding 97% (less than 3% errors) could be achieved by using only 5 000 SNPs when this method of selection was used; using other methods of selection this accuracy was only achieved when double the amount of SNPs (10 000) was used. In terms of SNP-wise imputation accuracy, accuracy estimates were lower for SNPs located on the chromosomal extremes and if the MAF of these SNPs was low. For chromosome 19, which was the chromosome with the worst mean imputation accuracy for most scenarios, SNPs located in the first (n=32), middle (n=42) and last (n=64) 1Mb of this chromosome, for example, had mean SNP-wise correlation-based accuracy measures of 0.640, 0.810 and 0.577. The difference in SNP-wise imputation accuracy moreover was 0.071 between SNPs in the highest (0.4<MAF≤0.5) and lowest MAF bins (0.01<MAF≤0.1); imputation accuracy was better for SNPs with higher MAF.

Results generated to achieve the final aim of this study are still preliminary and in the process of being analyzed. Preliminary results, however, shows strong correlations between conventionally-estimated EBVs and GEBVs, with the inclusion of genomic information being advantageous to breeding value estimation. The difference in GEBV accuracies estimated from true- versus imputed genotypes was small thus far, depending on the per animal imputation accuracy; the discrepancy is expected to be larger for animals with lower mean imputation accuracy.

Conclusion

The variation observed in genomic characteristics such as MAF and LD conformed to expectations and supported previous research suggesting that the SA Drakensberger is a composite breed with an admixed genome and heterogenous genomic architecture. This variation across the genome allowed variation in imputation accuracy between different chromosomes and genomic regions within chromosomes to be pre-empted. Genotype imputation is a valid genomic strategy for the SA Drakensberger breed and this study concluded that a genotyping panel consisting approximately 10 000 SNPs would suffice in achieving less than 3% imputation errors. Results presented further suggests that if such a panel were to be designed, that the SNPs considered for inclusion would have to be selected based on selection criteria, such as MAF and LD, specific to the SA Drakensberger breed. Considering that no Sanga-specific genotyping panel currently exists, it would be recommended that these SNPs be chosen from re-sequencing efforts, i.e. from a pool of SNPs that are identified as specific to the breed, and not necessarily from a pool of SNPs that are available on taurine- and/or indicine-derived genotyping platforms. The reason for this is that low MAF, because of ascertainment bias, was the most influential factor affecting  achievable imputation accuracy and therefore poses a concern. This study showed that it will be a valid strategy to integrate genotype imputation routinely into future genomic evaluation pipelines for the SA Drakensberger breed as imputation errors are expected to have a negligible effect on resulting GEBV accuracies. Finally, the inferences made from this study may be transferable to other Sanga breeds and may provide guidelines for consideration in future genomic endeavours for these breeds.

Popular Article

Genotype imputation as a genomic strategy for the South African Drakensberger beef breed by SF Lashmar, C Visser and FC Muchadeyi

The Drakensberger is a medium-framed breed of cattle with a sleek, black coat. It is believed to be one of South Africa’s oldest Sanga breeds and was developed from an ancestral population of cattle that was first sighted in 1659 in the Bredasdorp area of the Western Cape province. These cattle ancestors, also described as black in colour, belonged to native tribes and were crossbred with Dutch cattle of the Groningen breed, which were imported by European settlers in the 1700s. By this introduction of European Bos Taurus genetics, the development of the SA Drakensberger was initiated. The modern SA Drakensberger, as it is presently known, was however only recognized in 1947 when the SA Drakensberger Breeders’ Society was established. The breed therefore underwent a process of development that spanned centuries, whereby it withstood many harsh challenges in its history and this has led to the hardy breed it is today. Nicknamed the “profit breed”, the Drakensberger is both adapted and highly productive within SA’s beef producing environment and has a long history of diligent performance recording. In fact, it was the first breed to receive estimated breeding values (EBVs) using best linear unbiased prediction (BLUP) methodology, as performance testing was made compulsory to all breeders since 1980. Participation by Drakensberger breeders in SA Stud Book’s Logix Beef performance recording scheme is still at 100% today (SA Stud Book, 2017) and extensive pedigree records are available. Considering all of this, there has recently been interest in further enriching breed improvement strategies for the SA Drakensberger with genetic information in the form of genomic selection.

To implement genomic selection can significantly improve the efficiency of selection processes, and hence accelerate genetic progress, for the SA Drakensberger breed. This selection strategy, however, requires large numbers of animals to be “tested”, referred to as “genotyped”, for a high density of single nucleotide polymorphism markers (SNPs) in order to make reliable scientific deductions and to produce accurate genomic estimated breeding values (GEBVs) for farmers or breeders. From experience, international researchers have suggested 1 000 animals to be included in a training- or reference population to deduce the prediction equations that will be used in calculating GEBVs for selection candidates. Generating the amount of data to fulfill the number of genotyped animals necessary in the training population alone can become unfeasibly expensive, especially in developing countries, considering that the cost of genotyping an animal for about 150 000 SNPs is currently approximated at ZAR200 per animal. The cost of genotyping can, however, be alleviated by genotyping animals with SNP chips containing lower numbers of SNPs and “imputing” to higher density.

In statistical terms, imputation refers to the process of replacing missing data with substituted values. In the context of genomics, genotype imputation refers to a method of predicting SNP genotypes for SNPs that are either missing or were not physically genotyped. The genotypes are predicted based on patterns observed from a more complete data set of SNPs that are available for a group of animals that are representative of a specific breed. Consider for example that we have a young animal tested for 10 000 markers (which would be referred to as a “low-density SNP panel”) and the parents of this animal are tested for 100 000 markers (which would be referred to as a “high-density SNP panel”). Given the genetic relationship between the parents and the offspring, and the fact that these animals share large parts of the DNA, we can “impute” or infer the “missing” 90 000 markers for the young animal by making certain statistical assumptions using the principles of genetics. On a larger scale: if a “reference” population (consisting of older, high-impact animals with many offspring in the national herd) is genotyped for a high density of genetic markers (let’s say 150 000 SNPs) and a “test” population (younger, commercial animals in the national herd) is genotyped for a smaller subset of these SNPs (let’s say 50 000 SNPs), the 150 000-SNP genotype profile can be imputed for the “test” animals. The prerequisite is, however, that the animals in the reference- and test populations need to be related in some way, in other words they need to share underlying genetic patterns. These shared patterns can be used to fill in the gaps in SNP information. The imputed SNPs i.e. the 100 000 “missing” SNPs not included on the lower density panel, can however only be used in downstream application such as genomic selection if they were accurately imputed or assigned otherwise inaccurate scientific deductions will be made.

Imputation is now almost routinely included in genomic evaluation processes abroad because this methodology has been optimized, through trial-and-error and studying the factors influencing “imputability” of SNPs, for the most popular international breed. To be able to make use of this methodology within the South African beef industry, and more specifically for local breeds, requires a process of validation and this has not yet been performed for breeds such as the SA Drakensberger. The aim of the study was therefore to comprehensively evaluate genotype imputation for the SA Drakensberger breed so that it can be routinely applied in a GS pipeline.

The first step in the process of validation was to investigate the genomic characteristics of the breed. The genomic characteristics of SNPs have previously been shown to have an influence on the accuracy with which genotypes can be imputed. The genome of each animal is subdivided into different structures, called chromosomes, and on each of these chromosomes differences may furthermore exist between different DNA segments depending on the origin of these segments i.e. from which animal in the pedigree that part of DNA was inherited. As a result of the history of the SA Drakensberger, the genomes of animals belonging to this breed are expected to be composite i.e. containing genomic segments from both Bos taurus and Bos indicus. Certain parameters can provide more information on the SNPs within each of these segments and these include the minor allele frequency (MAF) and linkage disequilibrium (LD). The MAF gives an indication of the value of a specific SNP to the breed in question; if the MAF is high, it is an indication that both alleles of the SNP are present amongst the animals in the breed i.e. the SNP is informative. The LD provides an indication of the relationship between adjacent SNPs; if SNPs are in high LD, a “block” of SNPs can be inherited together and animals share larger parts of the genome with one another. This improves the ability of SNPs in these regions to be imputed. The software, Plink, was used to quantify these parameters on a per chromosome basis. Results showed that there was variation in these genomic characteristics between different chromosomes and this led us to expect differences in imputation accuracy between chromosomes.

The logical next step was to calculate the actual achievable imputation accuracy. The accuracy of imputation was calculated for imputation from several custom-derived low-density panels. To achieve this objective, different sets of SNPs were extracted from the SNP data available (150K SNP data) to mimic possible lower-density SNP panels. Panels containing 2 500, 5 000, 10 000, 20 000 and 50 000 SNPs were tested. The choice of SNPs to be included on each of these panels were based on certain SNP selection strategies i.e. different criteria were used to select the SNPs. The different strategies of selection included 1) selecting SNPs randomly, 2) selecting SNPs so that they were approximately evenly spaced, 3) selecting only SNPs with the highest MAF and 4) selecting SNPs based on a score combining its MAF and relationship to neighbouring SNPs (LD). Imputation was done using a software called FImpute and our findings suggested that a low-density SNP panel consisting of approximately 10 000 SNPs that were selected based on their MAF and LD information will be optimal. Using such a panel resulted in less than 3% imputation errors.

The final step was to determine the influence of mistakenly imputed SNPs on the accuracy of GEBVs and hence on genomic selection. The “single-step” approach to GS was tested using software called Mix99. Breeding values were calculated using 1) only pedigree information (traditional), 2) using true genotypic data (GEBV) and 3) using imputed genotypic data (imputed GEBV). These different breeding values were compared to determine whether imputation accuracy had an effect. Our preliminary findings suggest that the inclusion of genomic data is advantageous and that there is a minimal effect on GEBV accuracy estimates if imputation accuracy was good.

To conclude, results from this study indicated that imputation is a valid genomic strategy towards cost-effectively implementing GS for an indigenous breed such as the SA Drakensberger despite the uniqueness and complexity of its genome. The outcomes of this study may moreover be transferable to other Sanga breeds and may provide a set of guidelines for genomic studies requiring imputation in the future. Even though this study has shown that a more affordable lower-density panel can be developed from choosing SNPs with high MAF in indigenous breeds from currently available genotyping platforms, it would be invaluable for future genomic endeavours to develop a Sanga-specific panel using breed-specific SNPs identified from re-sequencing efforts.

Please contact the Primary Researcher if you need a copy of the comprehensive report of this project on :LindeD@arc.agric.za

Innovative management for beef productivity

Innovative management to increase beef productivity in South Africa : Phase II

Industry Sector: Cattle and Small Stock

Research focus area: Livestock production with global competitiveness

Research Institute: ARC – Animal Production Institute

Researcher: Ms SM Grobler MSc (agric) Animal Science

Research Team:

TitleInitialsSurnameHighest Qualification
DrM.M..ScholtzDSc (Agric) Animal Science
DrA.MaiwashePhD (Agric) Animal Science
MrP.J.J.BreytenbachMSc (Agric) Animal Science
DrJ.P.C.GreylingPhD (Agric) Animal Science
DrF.W.C.Neser
PhD (Agric) Animal Science

Final report approved: 2016

Aims of the project

  • To establish if synchronization can lead to an increase in the total mass of calves weaned from a limited calving season, most likely by decreasing the days to calving, but also by increasing number of calves born
  • To establish if breeding replacement heifers at 14 months have an economic advantage over breeding heifers at 26 months in term of reproductive performance
  • To establish the impact of the two different grazing strategies on veld condition, grass species composition and basal vegetation cover over time

Executive Summary

This report presents the final results of the combined first and second phase of the project: Innovative management for improved productivity: Beef.  The motivation for this project was that the South African beef market has changed with a need for livestock research and development to think in terms of a livestock systems approach.  This entails the combination of sound natural resource utilization, forage management and reproduction management to ensure a sustainable production enterprise over time through the allocation of limited resources.

When the project was planned it was clear that a period of 3 years was not enough to evaluate extremely valuable long term effects on herd life and veld condition; however, the RMRD-SA only fund projects for a maximum of 3 years and therefore a second application was approved to cover the remaining two-year project period.

South Africa is still a net importer of beef. Therefore, by increasing off take in the beef sector, South Africa can move towards self-sufficiency. With fertility being regarded as one of the main components influencing total beef herd efficiency, it is essential that the quoted calving percentage of 62% in the commercial beef sector of South Africa must be improved.  If the long calving seasons can be shortened and the calving percentage increased, more and heavier calves with a more uniform age can be weaned.  Cows that calve early also have a better chance of conceiving in the next breeding season and are generally seen as the more fertile animals

Development, production and quality of replacement heifers is a crucial component in the extensive beef production system.  In general, beef heifers are managed to calve for the first time at three years of age, but in some cases mating of heifers at one year of age have been advocated. 

All extensive beef production systems in South Africa are dependent on natural veld and it is well documented that veld condition have a huge influence on a number of beef production parameters. Studies conducted on natural veld have concentrated mainly on aspects that affect herd efficiency, including calving percentage, pre-weaning growth and supplementation of cows and calves.  However, none of the studies focused on the reproduction performance of beef cattle mated naturally after synchronization, heifer age at breeding and effect of grazing system on veld condition.

The aim of the study was to evaluate: the effect of estrous synchronization followed by natural mating on the calving percentage and calving distribution of multiparous beef cows and heifers; effect of breeding heifers at either 14 months or 26 months of age and the evaluation of a high utilized grazing system and controlled selective grazing on veld condition and animal performance.  The effects of climate on cow-calf production characteristics over time was also evaluated.

The study was conducted from 2009 to 2015 at the Roodeplaat experimental farm (REF) of the ARC-Animal Production Institute (25°34’11.27’’S; 28°22’05.36’’E) on 900 ha of natural rangeland described as Sourish Mixed Bushveld.  The experimental herd (n=92) was divided in four sub-herds consisting of 23 cows each at the beginning of the project in 2009.  It was ensured that the four sub-herds were as uniform as possible at the beginning of the project e.g. age, weight, previous number of calves. Within each sub-herd, 50% of the cows and heifers were synchronized prior to the commencement of the breeding season. Two sub-herds were subjected to high utilized grazing and two sub-herds were subjected to controlled selective grazing. The two grazing systems were related to the use of 30% or 60% of the available grass dry matter.  Half the heifers were mated at 14 months and the other half at 26 months.

Results from this study indicated that calving percentage and body condition score did not differ significantly (P=0.54) between cows that was either synchronized or not synchronized followed by natural mating.  However, estrous synchronization prior to natural mating did influence the average days to conception with synchronized cows calving earlier, except for 2012 in the calving season.  Over the six-year project period 15% more cows from the synchronized group conceived within 293 days after the onset of the breeding season. Calves from the synchronized cows weaned on average 5kg heavier than the cows that were not synchronized although this difference was not significant.

Conception rates of heifers mated at 26 months were significantly (P<0.05) higher than heifers mated at 14 months of age.  It would seem that it may be more viable to breed Bonsmara heifers in an extensive production system in the Sourish Mixed Bushveld region at 26 months of age for the first time.  Synchronization of 14 month old heifers did not improve conception rate over 14 month old heifers bred naturally.  However, the calving percentage of synchronized heifers bred at 26 months was 6% higher than the non-synchronized heifers.

Almost no veld condition change was recorded except for veld condition scores for both controlled selective grazing and high utilization grazing.  In addition, the results indicate a tendency that high utilization grazing improved veld condition score and grass species composition over that of controlled selective grazing, but the duration of the study is too short to make a definite conclusion on the effect of grazing strategy on veld condition.

It was also shown that grazing strategy did not have a significant influence on cow weight and calf growth over the six-year period, indicating that both grazing strategies are sustainable in the Sourish Mixed Bushveld if carrying capacity is adhered to. 

With the significant differences between years (P ≤ 0.05) for calving percentage, cow weight at calving, cow weight at weaning, calf birth weight, calf weaning weight and body condition score over the six-year observation period, the effect of seasonal temperature, relative humidity and rainfall is elucidated.  Forward stepwise regression procedures were performed to determine what climatic data were involved in cow and calf weight at birth and weaning as well as calving percentage.  In spite of the high standard errors (which were probably due to the small sample size), maximum relative humidity one month prior to the start of the breeding season, made a major contribution to explain calving percentage and minimum temperature within the last month of the 3 month breeding season, had a low negative correlation with calving percentage.   It can be speculated that high humidity in the study region (Sourish Mixed Bushveld) is an indication of warm and wet conditions, negatively impacting cow and bull comfort, leading to lower conception rates.  The negative correlation between minimum temperature within the last month of the breeding season and calving percentage may indicate that the cows were unable to cool down at night during the warmer summer months of the year, leading to lower conception rates and resorptions. The researcher acknowledge that the available herd size may be a limitation and that a bigger herd or sub-herds’ size combined with bigger land size could benefit the project outcome, possibly resulting in more significant differences and/or enhanced interpretation of results

Conferences

  1. Grobler, S.M., Scholtz, M.M. & J.P.C. Greyling, 2013.  Reproduction performance of beef cattle mated naturally after synchronization in the Central Bushveld Bioregion.  South African Society of Animal Science 47th Congress – University of the Free State, Bloemfontein, Free State Province 23-26 June 2013. Poster
  2. Grobler, S.M., Breytenbach, P.J.J. & M.M. Scholtz, 2013.  Effects of 2 grazing systems on veld in the Marikana Thornveld.  Grassland Society of Southern Africa.  48th annual Congress – Modimolle, Limpopo Province 15-19 July 2013.  Presentation
  3. Grobler, S.M., Scholtz, M.M., Neser, F.W.C., Greyling, J.P.C, Morey, L. & F. Calitz, 2016.  Reproductibve performance of extensively managed beef heifers mated at 14 months or 24 months in the Marikana Thornveld.  51th annual Congress – Stellenbosch, Western Cape Province 3 – 5 July 2016.  Poster

scientific articles

  1. Grobler, S.M., Scholtz, M.M., Greyling, J.P.C. & F.W.C. Neser, 2014.  Reproduction performance of beef cattle mated naturally following synchronization in the Central Bushveld bioregion of South Africa. S. Afr. J. Anim. Sci. 44: S70-S74
  2. Grobler, S.M., Scholtz, M.M., & Schwalbach, L.M.J. & J.P.C. Greyling, 2013.  Effect of synchronization on calving date following natural mating in beef cattle.  Appl. Anim. Husb. Rural Develop. 6:15-17

Popular Article

  • Grobler, S.M., Scholtz, M.M. & Breytenbach, P.J.J., 2014. Innovation = improved productivity. Red Meat/Rooi vleis. Agri Connect Pty (Ltd), Pretoria. Vol 5(1): 74-77

PhD Thesis submitted

  • Grobler, S.M., 2016. Alternative management systems to increase beef production under extensive conditions. PhD thesis. University of the Free State, Bloemfontein.

Popular Article

Will follow later

Jackal ecology on reserves and farms

Ecology and population dynamics of black-backed jackal (Canis mesomelas) on reserves and farms

Industry Sector: Cattle and Small Stock

Research focus area: Predation management

Research Institute: Centre for African Conservation Ecology, Nelson Mandela University

Researcher: Prof Graham Kerley

The Research Team

TitleInitialsSurnameHighest Qualification
DrLMinniePhD

Year of completion : 2017

Aims of the project

  • To determine dispersal direction between subpopulations
  • To compare demographic structures between subpopulatins
  • To resource use between subpopulations

Executive Summary

xxx.

Lethal carnivore management, aimed at reducing carnivore impacts, threatens the persistence of carnivores globally. The effects of killing carnivores will depend on their life histories and social structures. Smaller canids, like black-backed jackals (Canis mesomelas), are highly adaptable and display variable population-level responses to mortality sources, which may contribute to their success in fragmented landscapes. Jackals, the dominant predator of livestock in South Africa, are widely hunted to reduce this predation. This hunting is heterogeneous across the landscape, focussed on livestock and game farms, with nature reserves acting as refuges.

The aim of this research was to investigate the ecology and population dynamics of jackals in response to heterogeneous anthropogenic mortality. I hypothesized that the spatial variation in hunting results in the formation of a source-sink population structure, which contributes to the persistence of jackals. I addressed this hypothesis by evaluating two criteria, essential for the formation of a source-sink system in larger mammals.

Firstly, I confirm that hunting pressures result in the formation of distinct subpopulations with asymmetrical dispersal (i.e. compensatory immigration) from unhunted reserves to neighbouring hunted farms. Secondly, I show that jackal subpopulation display asynchronous demographics, with farm populations displaying a relatively younger age structure and an associated increase in reproductive output (i.e. compensatory reproduction). This confirms the formation of a hunting-induced source-sink system. Additionally, I show that jackals have a catholic diet, which confers a level of adaptability to direct (anthropogenic mortality, prey provisioning) and indirect (alteration in prey base) habitat modifications. This dietary flexibility allows jackals to obtain the appropriate resources to achieve reproductive condition. The relatively better body condition of younger jackals in sink habitats allows for compensatory reproduction which contributes to the success of jackals on hunted farms.

Based on my findings, I hypothesize that the compensatory life history responses of jackals to anthropogenic mortality may be ascribed to two interconnected mechanism. Dispersal is presumably driven by density-dependent interference competition, as dominant territorial pairs outcompete subordinates in high-density reserve areas, forcing them to disperse onto low-density farms (i.e. ideal despotic model). Additionally, farms likely represent attractive habitats, owing to a reduction in conspecifics and a concomitant increase in resource availability (including anthropogenic resource provisioning). Therefore, dispersing subordinates presumably select for farms which are perceived as good quality habitats, as the high risks of anthropogenic mortality cannot be perceived by dispersing individuals. This results in the formation of an attractive sink or ecological trap. These compensatory processes will continue to counter population management actions as long as recruitment from unmanaged areas persists. This hypothesis provides a conceptual framework for future research directions in understanding jackal persistence and management (i.e. specifically focussing on controlling dispersal) of jackal populations.

POPULAR ARTICLE

To follow soon

Please contact the Primary Researcher if you need a copy of the comprehensive report of this project – Graham Kerly on graham.kerley@mandela.ac.za

Jackal ecology on reserves and farms