Quality Deterioration of Hybrid Maize Seed During the Transfer from Seed Companies to Agro-Dealers in Uganda

TABLE OF CONTENTS

DEDICATION

ACKNOWLEDGEMENTS

ABSTRACT

TABLE OF CONTENTS

LIST OF FIGURES

LIST OFAPPENDICES

INTRODUCTION
1.1 Background
1.2 Problem statement
1.3 Justification
1.4 Objectives
1.4.1 Main objective
1.4.2 Specific objectives
1.5 Hypotheses

CHAPTER TWO LITERATURE REVIEW
2.1 Importance of maize
2.2 Constraints to maize grain production
2.4 Molecular markers in plant breeding
2.5 Single nucleotide polymorphism (SNP)
2.6 Phenotypic characterization
2.7 Genetic characterization

CHAPTER THREE MATERIALS AND METHODS
3.1 Study site
3.2 Plant materials and field evaluation
3.3 Data collection
3.3.1 Phenotyping
3.3.2 Genotyping
3.4 Statistical analyses
3.4.1 Phenotyping
3.4.2 Genotyping

CHAPTER FOUR RESULTS AND DISCUSSION
4.1. Phenotypic variation
4.1.1 The analysis of variance
4.1.2 Principal component analysis
4.1.3 Cluster analysis
4.2 Genetic variation
4.2.1 Allele frequency
4.2.2 Heterozygosity
4.2.3 Genetic distance
4.2.4 Cluster analysis

CHAPTER FIVE CONCLUSIONS AND RECOMMENDATIONS

REFERENCES

LIST OF APPENDICES

LIST OF TABLES

Table 1: Key participants in Uganda's formal seed system value chain as at 2010

Table 2: Bottlenecks affecting the seed value chain in Uganda

Table 3: Materials used in The study

Table 4: The phenotypic traits that were collected using visual scoring method

Table 5: Summary statistics of the phenotypic traits that showed significant differences

Table 6: Longe 6H eigenvalues and cumulative percentage of variation explained by the first four principal components

Table 7: Longe 7H eigenvalues and cumulative percentage of variation explained by the first four principal components

Table 8: Longe 9H eigenvalues and cumulative percentage of variation explained by the first four principal components

Table 9: Longe 10H eigenvalues and cumulative percentage of variation explained by the first four principal components

Table 10: Percentage of off-types within the field

Table 11: Codes of samples with their districts of origin as appearing in the genetic distance matrix

Table 12: Genetic similarity matrix of Longe 7H based on genetic distance

Table 13: Genetic similarity matrix of Longe 9H based on genetic distance

Table 14: Genetic similarity matrix of Longe 10H based on genetic distance

LIST OF FIGURES

Figure 1: Longe 10H Dendrogram, based on 33 phenotypic traits using Euclidean distance matrix and Average link method

Figure 2: Longe 9H Dendrogram, based on 33 phenotypic traits using Euclidean distance matrix and Average link method

Figure 3: Longe 6H Dendrogram, based on 33 phenotypic traits using Euclidean distance matrix and Average link method

Figure 4: Longe 7H Dendrogram, based on 33 phenotypic traits using Euclidean distance matrix and Average link method

Figure 5: Longe 7H dendrogram, based on the genetic distance using Roger’s similarity matrix and UPGMA clustering method 57 Figure 6: Longe 9H dendrogram, based on the genetic distance using Roger’s similarity matrix and UPGMA clustering method

Figure 7: Longe 10H dendrogram, based on the genetic distance using Roger’s similarity matrix and UPGMA clustering method.

LIST OFAPPENDICES

Appendix 1: Summary statistics of the phenotypic traits, MS, Average SED and Average LSD

Appendix 2: Principal Component Analysis (PCA) score plot for Longe 6H based on the 33 phenotypic traits

Appendix 3: Principal Component Analysis (PCA) score plot for Longe 7H based on the 33 phenotypic traits

Appendix 4: Principal Component Analysis (PCA) score plot for Longe 9H based on the 33 phenotypic traits

Appendix 5: Principal Component Analysis (PCA) score plot for Longe 10H based on the 33 phenotypic traits

Appendix 6: Summary statistics of 120 SNPs

Appendix 7: Summary of Heterozygosity and PIC of Longe 6H, 7H, 9H and 10H

DEDICATION

This thesis is dedicated to the Alidressey`s Family

ACKNOWLEDGEMENTS

I am very grateful to ALLAH almighty who gave me direction in the pursuit of this study and its accomplishment. In a special way, I wish sincerely to thank my supervisors, Dr. Sylvester Katuromunda and Dr. Godfrey Asea, for their support, guidance and constructive criticism which enabled me go through this research. I am greatly indebted to the unquantifiable contribution of Julius Pyton Sserumaga and also to Ssemaze for his technical support.

I also thank my sponsors INTRA-ACP Mobility project entitled: “Sharing Capacity to Build Quality Graduate Training in Agriculture in African Universities (SHARE)” who opened a window for me to undertake the studies at Makerere University, a chance that I highly regard as a great opportunity.

I am very grateful to my family members who have been patient with me, supporting and praying for me during this period of studies. My classmates and close friends, Sodeji Kpedetin Frejus Ariel, Tsindi Albert, Baguma Karubanga Julius, Nkurunziza Gelase, Weelar Charles, Yeannah George, Khalid Alsiddig and Agoyi Eric who have been so inspirational and have been my pillar of strength. Last but not least, I am very grateful to many others who are not cited here but who have contributed a lot to this study.

ABSTRACT

Seed market is becoming global and globalization is growing very fast. To compete favourably in this new global seed world, quality and cost are and will certainly continue to be the key issues. The presence of counterfeit seed in Uganda seed market has been reported. Studies need to be conducted to investigate at which level along the seed value chain the deterioration in quality occurs. The current study assessed the change in quality of hybrid maize seed as it is transferred from seed companies (when seed has been processed and packaged) to agro-dealers. Specific objectives of the study were to determine the levels of phenotypic and genotypic variation in hybrid maize seed from seed companies and agro- dealers. A total of 120 samples of four hybrid maize varieties (improved varieties most favored by farmers) used in this study were collected from agro-dealers in 15 districts and from 3 seed companies. These seed samples were planted in Alpha lattice design with two replications. Each entry was represented by two 5 metre long rows, with 30 cm in-row spacing and 75 cm inter-row spacing. Data was collected on 33 morphological traits and all the 120 samples were genotyped using 128 SNP markers.

The results of ANOVA revealed significant differences (P≤ 0.05) in 11 phenotypic traits out of 33 traits that were studied. Traits namely, T Anth, De Spi, T of S, Le of Ped, E Le, Anth BR, At B, Anth Col BG and Anth Col Si were significantly different. The PCA distinguished traits: In of An Col S, Anth Col BG, An Col Anth and Anth Col GeB as the important traits for distinguishing the maize hybrids. The major allele frequency ranged from 0.509 to 1.000. The PIC ranged from 0.00 to 0.38 for all the hybrids under the study. Longe 10H had the highest level of variation among the others with 46% heterogeneity, followed by Longe 7H (42%), Longe 9H (38%) and Longe 6H (35%). The genetic distance analysis showed that agro-dealers’ samples from Iganga, Masindi, Luwero, Soroti and Bukedea had high seed quality (genetically and phenotypically pure and similar to the seed companies’ seed) for all the hybrids under the study, whereas samples from Lira, Hoima, Mubende, Mityana, Gulu, Kiboga and Bugiri had high rate of variation/contamination among the agro-dealer samples. In addition, the study showed that the further the agro-dealers are from the seed companies, the more contaminated the seeds are. Therefore, further studies should be conducted to investigate the causes of variation in seed quality between seed lots at the seed company level, and to verify the agro-dealers’ behavior in the districts with high rates of seed contamination.

CHAPTER ONE INTRODUCTION

1.1 Background

Maize (Zea mays L., 2 n = 2 x = 20) is a monoic annual plant which descends from Maydeas tribe and the grass family Gramineae. However, the most accepted diverse theory of the origin and evolution of maize is that, maize’s center of origin is located in Mesoamerica, primarily Mexico and the Caribbean (Verheye, 2010). It was domesticated from the wild grass Teosinte (Zea mays sp mexicana/ Euchlena mexicana) at least 6000-6250 years ago. Maize was brought to Africa and later spread to other tropical countries, mainly by the Portuguese and Arab explorers (Verheye, 2010).

Maize is one of the most important cereal crops in the world together with wheat and rice (Devi et al., 2013), and the second most important crop in terms of quantity production after sugarcane (FAOSTAT, 2012). In 2013, the United States of America was the world’s largest producer of maize. It produced 273,832 metric tons (MT), followed by China (205,600 MT), Brazil (81,000 MT) and South Africa (12,200 MT) (Statista, 2013). The great multiplicity of environments and conditions have shaped the basis for the development of maize varieties that are well adapted to harsh conditions of soil and climate as well as to biotic stresses (Devi et al., 2013).

Maize provides at least 30% of the food calories to more than 4.5 billion people in 94 developing countries. It is a staple food for 900 million consumers in Africa, Asia and Latin America (Shiferaw et al., 2011). Currently, maize is produced on nearly 100 million hectares in 125 developing countries and is among the three most broadly grown crops in 75 of those countries (FAOSTAT, 2010). It plays an important role in the livelihoods of many poor farmers. This is because about 67% of the total maize production in the developing world comes from low and middle income countries (Shiferaw et al., 2011). Statistics show that, of the 23 countries with a high per capita consumption of white maize, 16 are in sub-Saharan Africa (SSA) (Khoza, 2012). It supports millions of livelihoods and livestock in SSA in terms of food and feed, respectively.

Maize can be utilized in many ways in comparison to other cereals. Almost, all plant parts of maize have economic value. For example, maize stems after harvest (stover) can be used as livestock feed, mulching materials, or fuel wood for cooking. In SSA, maize is used mainly for human consumption, while in industrialized countries it is used as livestock feed and as raw material for industrial products such as plastics, fabrics, adhesives, biofuel (ethanol). It is an important source of carbohydrate, protein, iron, vitamin B and minerals (Khoza, 2012).

In Uganda, maize is the most important cereal crop followed by sorghum and millet in terms of area under cultivation, production and human consumption. An average of 1.5 tons of maize per hectare is produced in Uganda (Agona and Muyinza, 2001). It provides 9.2% of the total calorie intake in Uganda, 34.1% in Tanzania, and 35.2% in Kenya as the highest in East Africa (Gichuru, 2013).

Maize breeding efforts have included addressing yield deterioration as a consequence of low soil fertility, low fertilizer use, poor agronomic performance, and high susceptibility to diseases such as maize streak virus, turcicum leaf blight and grey leaf spot. In 2002, the Uganda cereals research programme released three maize varieties, namely Longe 6H, Longe 7H and Longe 8H. In 2009, Yara 41, Yara 42, Longe 9H, Longe 10H and Longe 11H were also released (Barnett et al., 2011). Future research efforts will target resistance to diseases and pests, drought and striga. Since the release of Longe 6H, Longe 7H, Longe 9H and Longe 10H varieties, their purity and productivity has decreased as a result of contamination during multiplication (seed production), processing and marketing. The reduction in purity and productivity of seeds of these varieties has in turn affected the grain yields obtained by farmers, hence causing farmers to lose confidence in the seeds of improved varieties on the market.

1.2 Problem statement

Uganda’s seed sector compared with other countries in SSA is characterized by the fact that the government recognizes both the formal and informal seed production and supply systems in its policies and programmes. The formal seed production and supply system produces high quality seeds of genetically improved varieties. It aims at enabling farmers access seed of high quality, that is, genetically, physiologically (germination potential and vigour) and analytically pure, as well as being free from seed-borne pathogens. Production of high quality seeds is accomplished by regular and rigorous monitoring and supervision of seed crop in the field, and the seed after harvest during processing and packaging, as well as confirmation of quality through laboratory tests before sale to farmers. Thus, the seeds are expected to have better yield potential than seeds from the informal seed sector (i.e., farmer-saved seed and that produced under the community based seed systems).

The formal seed production and supply system is a well-defined system comprising of variety development, seed multiplication, marketing and distribution, and quality control. These components are handled by different organizations, which include government institutions/departments, research/plant breeding institutions, seed companies, seed growers, seed processors and seed distributors/stockists/seed sellers. Currently, Uganda has 23 registered seed companies that are active in the seed market. They are involved in the multiplication, processing and marketing of seeds of cereal, legume, horticultural, oil and forage crops in the domestic and regional markets. These companies receive or buy breeder seed from NARO institutes (such as NaCRRI Namulonge and NaSARRI Serere), that has been approved for multiplication and released by the National Variety Release Committee. Using contract seed growers, seed companies multiply the breeder seed into certified seed, and process it in their own or hired factories. Seed companies then market their seed, mainly through a network of seed distributors and stockists who in turn sell seed to individual farmers and farmer groups (Mubangizi et al. 2012).

Despite all the efforts taken to ensure that high quality seed is available to farmers, Uganda’s seed sector is characterized by counterfeit seeds, which leads to inability of farmers to exploit the full potential of improved varieties. Barnett et al. (2011) reported that 40% of seed available on the market in Uganda is adulterated. Adulterated/counterfeit seeds can be considered to be seeds of poor quality, with low germination capacity, mixtures of various varieties, seed which has lost vigour due to overstaying in storage, and grain packaged in reused seed containers. Diluting and counterfeiting of seed can happen at all levels along the production and supply (value) chain. For example, seed companies typically sell seed packaged in larger bags, but the bags are opened and the seed is often split up at the retail/agro-dealer level into smaller units, because that's what farmers can afford to buy. This creates an opportunity for the agro-dealers to adulterate the seeds. The presence of adulterated seeds on the market is reducing farmers’ confidence in the seed supply chain (as a result of the poor performance), and in turn negatively affecting the adoption of improved varieties.

In order for farmers to obtain higher grain yields as reported by breeders and researchers, it is important that the purity of seed be maintained at all levels of the seed system, right from breeder seed up to certified seed supplied on the market for farmers to plant. Previous study on seed purity and varietal identity in the maize seed value chain has shown that variation (loss of purity) occurs between the breeder seed and the seed companies’ seed. The current study is a follow up with the objective of determining the level of phenotypic and genetic variation in hybrid maize seed as it moves from seed companies (when seed has been processed and packaged) to agro-dealers.

1.3 Justification

The development of seed supply systems is a precondition for increasing food production, improving farmers’ incomes through increased productivity, alleviating poverty and ensuring food security (FNSU-UDS, 2011). However, a vibrant seed system is critical for increased production and productivity of any agricultural system.

Maize is an important staple food in Uganda, and dominates overall seed production, with almost 7000 tons produced in 2009 (Uganda AGRA-PASS, 2010). Maize has a well- developed seed system, and is grown widely throughout the country. Uganda’s maize seed market is small, but is growing rapidly at an estimated rate of 5 to 10 percent annually (Larson and Mbowa, 2004). Even so, adoption rates remain low relative to many neighboring countries. However, there are high incidences of low/poor seed quality due to inadequate inspection and monitoring of seed dealers by Uganda Seed Companies System (USCS) caused by lack of capacity/personnel and logistics; Lack of operational standards/guidelines by seed companies and enforcement of seed regulations after privatization; Lack of enough seed experts and laboratory testing equipment; Low or poor supervision of seed producers; Unscrupulous seed dealers; and counterfeit seed and out-crossing (PELUM, 2012). The morphological and molecular characterization of maize genetic material along the seed production and supply chain would enable us to determine the stage at which variation occurs, which is important for crop improvement, and would enhance the quality of maize seed in Uganda.

1.4 Objectives

1.4.1 Main objective

To contribute to the improvement of hybrid maize seed production by determining the genetic and species purity of seed from seed companies and agro-dealers

1.4.2 Specific objectives

1. To determine the level of phenotypic variation in hybrid maize seed from seed companies and agro-dealers

2. To determine the level of genotypic variation in hybrid maize seed from seed companies and agro-dealers

1.5 Hypotheses

1. The phenotypic variation might occur in hybrid maize seed during transfer from seed companies to agro-dealers.

2. The genotypic variation might occur in hybrid maize seed during transfer from seed companies to agro-dealers.

CHAPTER TWO LITERATURE REVIEW

2.1 Importance of maize

Globally, maize (Zea mays L.) is the second cereal crop after wheat, with about 600 million MT produced annually (Nyombayire et al., 2011). In SSA, maize is considered a very important food and feed crop. Farmers harvest 25 million hectares of maize per annum, producing about 35 million MT of grain. This accounts for 40% of the region’s cereal production. Most of the grain (90%) is consumed by humans (Bänziger, 2001). Out of 23 countries in the world with the highest per capita consumption of maize as food, 16 are in SSA. Maize provides 50% of the calories in diets in southern Africa, 30% in eastern Africa, and about 15% in west and central Africa. More than 300 million Africans rely on maize as their staple food. The grains are rich in vitamins A, C and E, carbohydrates, essential minerals and contain 9% protein. Maize grain is also rich in dietary fiber (Dao, 2013).

In Uganda, maize is cultivated on about 1.5 million hectares of land, and it’s the third most cultivated crop after banana and beans. Maize has gradually become a staple food, substituting crops like sorghum, millet, cassava and banana (Ahmed, 2012). An average yield of 1.5 MT of maize per hectare is produced. Most of this maize, in addition to being eaten directly as food, supports the local brewery industry, where its flour is fermented to produce local beverages. The residues such as stover and bran serve as livestock feeds (Agona and Muyinza, 2001).

2.2 Constraints to maize grain production

Maize production in Africa was estimated at 2.1 MT/ha in 2012 which was far less than the global (4.9 MT/ha), European (5.1 MT/ha), Asian (5.0 MT/ha) and American yields (6.3 MT/ha) (FAOSTAT, 2012). Numerous factors justify the low yield of maize in Africa including abiotic and biotic constraints, and socio-economic factors. In a given season, crops can be ruined by one or a number of constraints.

2.2.1 Drought

Drought is usually as a result of inadequate or poorly distributed rainfall which decreases maize output across the region. Severe droughts periodically cause deficits in cereal production leading to food crisis particularly in the region (Dao, 2013). The Food and Agricultural Organization (FAO) estimated that SSA is the most severely affected region where almost half of the land surface is exposed to a high risk of meteorological drought (Ribaut et al., 2004). Drought affects all the agro-ecological zones even the high-yield potential regions where the crop may be affected by mid- and late-season droughts. Drought affects maize yields by limiting season length and through random stress that can occur at any time during the cropping cycle (Dao, 2013). It is affecting 21% of the maize area and reducing yields by 33%, on average. Droughts in 1992 and 2002 reduced maize production in the southern Africa region by 50% (Semagn et al., 2014). The negative impact of drought may grow as the threat of climate change becomes a reality.

2.2.2 Soil nutrient deficiency

Nutrient reduction and soil fertility decline are extensive in smallholder farming systems, as many farmers cannot afford or do not have access to organic and inorganic fertilizers. Soil fertility deterioration is generally related to increased population pressure, especially in areas with fragile ecosystems, such as SSA (Dao, 2013). Rijsberman (2006) observed that drought stress and poor soil fertility will keep on causing negative effects to agricultural production in the coming years, mostly in Asia and Africa.

2.2.3 Parasitic weeds

Striga hermonthica is one of the most important biotic factors limiting maize grain yield in the savanna zones of SSA. Yield losses attributable to this obligate hemi parasite may range from 10 to 100% depending on the genotype grown, climatic conditions, soil fertility status and levels of infestation (Akaogu et al., 2013).

2.2.4 Socioeconomic factors

Among the socioeconomic and political boundaries on food production are public policies and investments that are biased against poor farmers and consumers, women and less-favoured areas; inadequate infrastructure; inequitable access to land and other critical resources; poorly functioning and poorly integrated markets; and lack of access to credit and technical assistance. The few facilities and services provided are most of the time primarily made available for men than women farmers although the latter produce about 75% of the domestically grown food in SSA (Dao, 2013).

2.2.5 Pests and diseases

Maize production faces a number of constraints, including low yields due to pests and diseases. Herbivorous insects are responsible for destroying 1.5% of the world's total crop production annually (Mohamed, 2013). The most serious insect pests of maize in Africa are stem borers (especially Busseola fusca, Eldana saccharina, Sesamia calamistis and Chilo partellus) which are responsible for 34-43% yield losses (Bamaiyi and Oniemayin, 2011; Mohamed, 2013); cutworms (Agrotis spp.), cob borer (Mussidia nigrivenella), cotton bollworm (Helicoverpa armigera), armyworm (Spodoptera exempta), leafhoppers (Cicadulina spp.) and variegated grasshopper (Zonocerus variegatus) (Babu-Apraku et al., 2012).

The most important fungal diseases of maize are Grey leaf spot (Cercospora zea-maydis) which is responsible of over 70% yield losses in USA (Benson et al., 2015), rots affecting female inflorescences (Fusarium spp. and other fungi), the stalk-rot complex (Diplodia maydis, Fusarium moniliforme, Macrophomina phaseoli and Pythium aphanidermatum) and leaf blights (Exserohilum turcicum and Bipolaris maydis), downy mildew (Peronosclerospora sorghi), maize smut (Ustilago maydis) and rusts (Puccinia sorghi and Puccinia polysora). For the case of viral diseases, maize streak virus (MSV) is the most important, and is restricted to Africa and may cause 100% yield loss (Babu-Apraku et al., 2012), followed by maize dwarf mosaic virus (MDMV), sugarcane mosaic virus (SCMV) and maize chlorotic mottle virus (MCMV). Common storage pests of maize are grain moths (Sitotroga cerealella and Ephestia cautella), grain weevils (Sitophilus spp.) and the larger grain borer (Prostephanus truncatus).

2.2.6 Seed quality

Elepu and Dalipagic (2014) defined value chain as “the full range of activities which are required to bring a product or service from conception, through the different phases of production (involving a combination of physical transformation and the input of various producer services), delivery to final customers, and final disposal after use”. The chain actors who actually transact seed as it moves through the value chain include input suppliers (e.g., seed), farmers, traders, processors, transporters, wholesalers, retailers and final consumers. Table 1 summarizes the actors in the Uganda’s formal seed system value chain. However, the study undertaken by CIMMYT and IITA in 2007 indicated that Uganda has been unable to take full advantage of the recent advances in seed sector development mainly because of institutional bottlenecks affecting the seed value chain. Table 2 summarizes the major bottlenecks affecting the seed value chain in Uganda (Ssebuliba, 2010).

Table 1: Key participants in Uganda's formal seed system value chain as at 2010

Abbildung in dieser Leseprobe nicht enthalten

Source : Uganda AGRA-PASS (2010)

KEY: MNCs = Multi-National Companies, SMEs = Small and Medium Enterprises, NGOs = Non-Governmental Organizations, CBOs = Community Based Organizations

Table 2: Bottlenecks affecting the seed value chain in Uganda

Abbildung in dieser Leseprobe nicht enthalten

Source: Uganda Seed Trade Association (USTA) (2010)

Salgado (2006) observed that seed quality affects directly the crop productivity, and it can be emphasized that it is a combination of genetic, physiological and sanitary attributes. Many standards are adopted to maintain the genetic purity of maize hybrid seeds, such as the isolation of the seed production fields, removal of female parental tassel, and cleaning the harvesting and processing machinery. In fields for production of maize hybrid seed, the main source of genetic contamination is self-pollination of the female parent due to incomplete removal tassel. This contamination increases the endogamic levels, reducing the genetic and physiological quality of seeds that consequently decreases crop productivity. In addition, the parental samples might not be pure and may comprise mixtures of hybrid seeds. The hybrid seeds may get mixed by hand or machine during processing (Weishi et al., 2012). As we follow the seed value chain, the agro-dealers themselves may sell adulterated seed or mix the improved seed with grain hoping to increase seed quantities and make more money. However, seed is the key to agricultural production and productivity and under optimum conditions account for up to 40% of yields. Any strategy to improve agriculture must, therefore, be anchored in a vibrant and effective seed value chain that produces, evaluates, multiplies and distributes high quality seeds to farmers.

2.3 Uganda’s Seed System

The seed system in Uganda is composed of the formal and the informal seed production and supply systems. The formal seed system is responsible for the production of improved and certified seeds. Seed production is characterized by a structured system of multiplication, distribution, marketing and quality control. The formal seed system provides about 10% of improved and certified seed used in the country, while the informal seed system provides the rest (90-95%) though this seed is not certified and thus its quality is unknown (Barnett et al., 2011). The number of registered seed companies are 22 (Ssebuliba, 2010). The seed market is dominated by NGOs and international agencies for relief interventions in northern Uganda and neighboring countries. Maize dominated the seed production and accounted for almost 80% of the approximately 9000 MT of certified seed that was produced in 2009 (Uganda AGRA-PASS, 2010). Government policy is highly supportive of private sector participation in seed production.

Uganda is member of the Organization for Economic Co-operation and development (OECD) seed schemes. The seeds and plant Act of 2006 regulates the promotion, regulation and control of plant breeding and variety release, multiplication, conditioning, marketing, importing and quality assurance of seeds and other planting materials, and for other related matters. However, there are a number of laws, regulations, administrative and technical procedures that regulate seed production and distribution in Uganda. These include: The Agricultural Chemicals (control) Act 2006, The Adulteration of Produce Act, Cap. 27; The Cotton Development Act, Cap. 30; The Uganda Coffee Development Authority Act, Cap. 325; The Uganda National Bureau of Standards Act 1983; The Export Promotions Board Statute, 1996; The National Environment Act, Cap. 153 and The National Agricultural Research Act, 2005 (Barnett et al., 2011).

Limited access to capital to invest in infrastructure and facilities limits the production capacity of seed companies. As a result of inadequate staffing and under funding of the national seed certification service (NSCS), Uganda seed system lacks the means and authority to regulate seed quality effectively and to regulate the market for other agro-inputs (Uganda AGRA- PASS, 2010). In addition, breeder and foundation seed quality remains questionable and there has been poor enforcement of royalty agreements, and adulterated inputs in the market.

2.4 Molecular markers in plant breeding

The application of molecular markers in plant breeding has been increased by advances in two key areas. The first is improvement of molecular marker technologies and the second is improvement of our ability to associate markers with traits (Vogel, 2010). Molecular markers are progressively being adopted by researchers involved in crop improvement as an effective and appropriate tool for addressing several basic and applied research areas relevant to agricultural production systems (Prasanna and Hoisington, 2003). In the last two decades, multiple generations of DNA detection technologies were developed. These include: restriction fragment length polymorphisms (RFLPs), simple sequence repeats (SSRs), amplified fragment length polymorphisms (AFLP) and single nucleotide polymorphisms (SNPs) (Fraley, 2009).

DNA fingerprinting and genetic diversity analysis using molecular markers is of significant importance in effective management of germplasm collections (Prasanna and Hoisington, 2003). Emphasis is also being put on comprehensive analysis of genetic diversity in breeding materials of major crops (Mohammadi and Prasanna, 2003). DNA-based markers are also being used to discover and utilize the evolutionary relationships between numerous genera within a family (e.g., the grass family, Poaceae), and various species within a genus. Genetic mapping of members of the agriculturally-important grasses, including rice, wheat, maize, sorghum and sugarcane with common DNA probes has exposed remarkable conservation of gene content and gene order (Devos and Gale, 2000). Advances in marker technology has played an important role in improving our ability to associate markers with traits.

In maize breeding, using molecular markers is particularly helpful for maintenance and broadening of the genetic base of the elite germplasm, assignment of lines to heterotic groups, selection of appropriate parental lines for hybrid combinations and generation of segregating progenies with maximum genetic variability for further selection (Prasanna and Hoisington, 2003). Progress in genomics has led to the identification of numerous DNA markers in maize during the last few decades, as well as thousands of mapped microsatellite or simple sequence repeat (SSR) markers, and more recently, single nucleotide polymorphisms (SNPs) and insertion-deletion (InDel) markers. In addition to the SSRs and SNPs, a large number of genes controlling various aspects of plant development, biotic and abiotic stress resistance, quality characters have been copied and characterized in maize, which are excellent assets for molecular marker-assisted breeding (Prasanna et al., 2010).

Dekkers and Hospital (2002) suggested two applications of molecular markers in breeding programmes. In the first approach, markers are used to improve a population through selection. In what is referred to as marker-assisted selection (MAS), phenotypic information is combined with genotypic information to identify the superior germplasm for commercial production. In the second approach, molecular markers can be used in introgression programmes. The goal of such breeding programmes is to transfer a desired trait or gene from one population into the background of an elite line/variety. Many of the traits that are under selection in breeding programmes are complex quantitative traits which are controlled by many genes plus environmental factors (Dekkers and Hospital, 2002). Molecular marker breeding approaches provide improved selection of breeding populations for these traits. In commercial breeding programmes, breeders develop a selection model for a breeding population and genotype the progeny with particular molecular markers (Vogel, 2010).

However, molecular markers have been used to identify and characterize quantitative trait loci (QTL) associated with diverse traits in maize including grain yield, characters concerned with domestication, environmental adaptation, disease and insect pest resistance, and drought and heat stress tolerance (Prasanna and Hoisington, 2003).

2.5 Single nucleotide polymorphism (SNP)

A SNP is a single change in the sequence of a section of DNA. It may come about as a result of a substitution of one nucleotide for another at the polymorphic site. A SNP can also be a single base insertion or deletion variant referred to as an indel. The majority of SNPs are biallelic; however, they can also be tri- or tetra-allelic. Tri-allelic SNPs involve the presence of three different nucleotides for the specific SNP whilst tetra-allelic involve four nucleotides (Ndhlela, 2012). Two main strategies have been employed to identify SNPs in plants. These are utilization of expressed sequence tag (EST) sequence information to direct targeted amplicon resequencing, and next generation sequencing (NGS) technologies coupled or not to genome complexity reduction methods (Grattapaglia et al., 2011).

Molecular markers based on SNPs are many, evenly dispersed throughout the entire genome and adequate to discriminate individuals in a population. According to Parida et al. (2012), SNP markers have gained significant importance in plant genetics and breeding because of their excellent genetic attributes and suitability for genetic diversity analysis and evolutionary relationships, understanding of population substructure, detection of genome-wide linkage disequilibrium, and association mapping of genes controlling complex phenotypic traits. SNP markers have become the most widely used markers because they target single nucleotide differences between genotypes, showing more polymorphism compared to other types of markers (Ndhlela, 2012). According to Van Inghelandt et al. (2010), SSR markers have been the most widely used DNA marker type to characterize germplasm collections of crops because of their easy use, relatively low price, and high degree of polymorphism provided by the large number of alleles per locus. In comparison, SNP markers received high attention because they occur at much higher frequency in the genome than SSRs and their genotyping can be easily automated. However, maize has a relatively high frequency of SNPs (Tenaillon et al., 2001). Hamblin et al. (2007) compared 89 SSRs and 847 SNPs for characterization of 259 maize inbred lines and found that SSRs performed better at clustering germplasm into populations than SNPs. They suggested that large numbers of SNP loci will be required to replace highly polymorphic SSRs in the study of diversity and relatedness. Large numbers of SNP markers have been identified through the whole genome sequencing of maize inbreds, sequencing of genomic fractions with reduced complexity (i.e., through the elimination of highly repeated DNA sequences) or transcriptome sequencing (Buckler et al., 2011).

2.6 Phenotypic characterization

A maize breeding programme relies heavily on the knowledge of breeding materials genetic relationships of the interested traits. This helps in identification of contrasting traits and avoids work duplication. Genetic relationship studies are achieved by the use of morphological traits, biochemical and molecular characterization markers (Bucheyeki, 2012).

Morphological markers characterization is preferred because, it is cheap and simple method used to determine the genetic relationships among species. Furthermore, several studies have been reported to utilize this method and find reliable and useful information in various crops.

Fetahu et al. (2012) studied the genetic and phenotypic diversity among some common bean landraces and successfully evaluated 15 common beans and three accessions of cow peas in Bulgaria. Bucheyeki et al. (2008) employed morphological characterization to study 11 qualitative and 26 quantitative traits of 37 sorghum landraces collected mainly from Tanzania. In maize, several researchers have utilized phenotypic markers to study genetic relationships among germplasm. For instance, Gabriel et al. (2009) used phenotypic markers to create genetic relationships information among maize forage landraces in Brazil. Ruiz and Alvarez (2001) employed phenotypic markers to study the genetic relationships of 100 landraces in Spain basing on twenty-two morphological traits, and seventeen ecological variables (climatic, edaphic and topographic) associated with the collection site. However, morphological characters are highly affected by environment (Cadee, 2000), hence the need for the employment of genetic markers.

2.7 Genetic characterization

Genetic characterization refers to the detection of variation as a result of differences in either DNA sequences or specific genes or modifying factors (Vicente e t al., 2005). Genetic characterization of elite material has received special attention since the late 1980s because of its strategic interest for breeders’ rights protection, analyses of homogeneous material (inbreds) are easier than using unfixed material, and it’s easy to determine regions of the genome QTL that contribute to the variation of traits of agronomic importance for many plant species (Charcosset and Moreau, 2004). Several studies have been reported to use genetic characterization method and found reliable and useful information in various crops.

Kumar et al. (2014) studied the genetic diversity and relationships among 38 genotypes of Plantago representing seven species using phenotypic and molecular markers in India. The study showed that molecular markers are highly discriminatory, more precise and more reproducible than some morphological traits used to estimate genetic relatedness among Plantago genotypes. Amankwaah (2012) used morphological descriptors and molecular markers to assess similarity of sweet potato of Ghanaian released and elite varieties compared with putative ramets in Ghana. Semagn et al. (2014) investigated the degree of genetic differences, patterns of relationships, and population structure among 218 diverse open pollinated varieties (OPVs) widely used in southern and eastern Africa. Lanes et al. (2014) studied the genetic relationships among 90 maize inbreds that were derived from tropical hybrids and populations. A total of 1229 informative SNPs and 1749 haplotypes within 327 loci were used to estimate the genetic diversity, population structure and familial relatedness in maize (Yan et al., 2009). These studies indicated that, genetic characterization with molecular technologies offers an enhanced greater power for detecting diversity (including genotypes and genes) than do phenotypic methods (Vicente et al., 2005).

CHAPTER THREE MATERIALS AND METHODS

3.1 Study site

The study was conducted at National Crops Resources Research Institute (NaCRRI), Namulonge. It is located within the bimodal rainfall region. It lies at 00 32'’ N of the Equator and 32⁰ 37'’ E, and at an altitude of 1200 m above sea level. It is located 27 km north of Kampala, the capital city of Uganda. It has a tropical wet and mild dry climate, and receives annual rainfall ranging between 800 and 1100 mm, with slightly humid conditions (average 65%), average annual temperature of 22⁰ C, and annual minimum and maximum temperatures of 16 and 28⁰ C, respectively. The soils are dark, reddish brown, sandy loam, with a pH range of 5.5 to 6.2.

3.2 Plant materials and field evaluation

A total of 49 seed samples of four hybrid maize varieties popular with farmers were collected from different agro-dealers in 15 districts and three seed companies. These together with 71 samples which had already been collected during the previous study from the same 15 districts generated a total of 120 samples for this study (Table 3). Seeds were sown in an Alfa Lattice design with two replications. Each entry was represented by 5 m long rows, with 30 cm in-row spacing and 75 cm inter row spacing, and two seeds were planted on each hill. This spacing was selected to eliminate any interplant competition and thereby allow plant traits to be clearly expressed (Lafarge and Hammer, 2002).

Table 3: Materials used in the study

Abbildung in dieser Leseprobe nicht enthalten

3.3 Data collection

3.3.1 Phenotyping

Data was recorded from 10 randomly selected plants of each sample basis on 33 traits, using visual scoring methods developed by CIMMYT as illustrated in Table 4. Each colouration trait was scored on a 1 to 9 scale, with one being absent or very week, three-weak, five-medium, seven-strong and nine-very strong. The angle between leaf blade and stem and the angle between main axis and lateral branches in the tassel were scored on a 1 to 9 scale with one being very small, three-small, five-medium, seven-large and nine-very large. Attitude of blade and attitude of lateral branches were scored on a 1 to 9 scale with one being straight, three- slightly recurved, five-recurved, seven-strongly recurved and nine-very strongly recurved. Degree of zig zag was scored on a 1 to 3 scale, with one being absent or very slight, two-slight and three-strong. Density of spikelets was scored on a 3 to 7 scale, with three being lax, five- medium and seven-dense. Ear shape was scored on a 1 to 3 scale, with one being conical, two- conical-cylindrical and three-cylindrical. Ear length and ear diameter were scored on a 1 to 9 scale, with one being very short, three-short, five-medium, seven for large and nine for very large. Length of husk off the tip of the ear, Length of main axis above lowest side branch, Length of main axis above upper side branch and Length of peduncle were scored on a 1 to 9 scale, with one being very short, three-short, five-medium, seven-long and nine-very long. Shape of tip was scored on a 1 to 5 scale, with one being pointed, two-pointed or round, three- round, four-round and spatulate and five-spatulate. Time to anthesis (50% of plants having tassels) and time to silk emergence (50% of plants with silks) were scored on a 1 to 7 scale, with 1 being very early and 7 for being very late. Type of grain was scored on a 1 to 7 scale, with one being flint, two-flint-like, three-intermediate, four-dent-like, five-dent, six-sweet and seven-pop. Width of blade was scored on a 1 to 9 scale, with one being very narrow, three- narrow, five-medium, seven-wide and nine-very wide.

Table 4: The phenotypic traits that were collected using visual scoring method

Abbildung in dieser Leseprobe nicht enthalten

3.3.2 Genotyping

DNA isolation, quality and quantity check. Leaf samples for DNA isolation were collected from 10-14 day-old seedlings and the extraction and purification of the genomic DNA from each accession was carried out using CTAB method from lyophilized leaf tissue (Hoisington et al., 1998; Murray and Thompson, 1980). DNA quality of each sample was assessed by electrophoresing the DNA. All the DNA samples were uniformly diluted to have a final concentration of 10 ng/μl. Then the samples were sent to LGC genomics for genotyping. The SNPs markers were used following the KASPar (K Bioscience Competitive Allele-Specific Polymerase chain reaction assay) SNP genotyping system (http://www.ksre.ksu.edu/igenomics/doc1363.ashx).

3.4 Statistical analyses

3.4.1 Phenotyping

For each trait, an analysis of variance (ANOVA) was conducted using the ASREML procedure (Magorokosho, 2006). Principal Component Analysis (PCA) was performed on the phenotypic correlation matrix of the adjusted means of the 33 descriptors so as to estimate the variation between the traits (Ignjatović-Micić et al., 2013). A dendrogram was constructed using Ward method (Ward, 1963) to provide a general visualization of the relationship between hybrids based on 33 traits. All the statistical analyses were performed using GenStat 14th Edition.

3.4.2 Genotyping

Allele frequency, unbiased estimation of gene diversity, observed heterozygosity and polymorphism information content (PIC) value were calculated using Power Marker software V3-25 (Liu and Muse, 2005). The PIC value (Botstein et al. 1980) was used to inter the relative value of each marker with respect to the amount of polymorphism revealed. Heterozygosity and unbiased gene diversity were calculated to quantify the genetic variation in the maize hybrids sampled. The genetic distance between samples was computed using the Roger’s genetic distance (Rogers, 1972) with Power Marker software V3-25. A dendrogram was constructed, in cluster analysis from the Roger’s genetic distance matrix using UPGMA with Power Marker and the resulting trees were visualized using MEGA version 5.2.2 (Tamura et al., 2011).

CHAPTER FOUR RESULTS AND DISCUSSION

4.1. Phenotypic variation

4.1.1 The analysis of variance

In this study, 33 phenotypic traits (Table 4, Appendix 1) and SNPs markers were used to characterize a set of four maize varieties (Longe 10H, 9H, 7H and 6H) collected from agro- dealers in 15 districts and three seed companies in Uganda. The analysis of variance revealed no significant differences within the hybrids for all the traits except for, Time of anthesis (T Anth), density of spiklets (De Spi), length of main axis above lowest side branch (Le MALB), ear length (E Le), length of peduncle (Le of Ped) and time of silk emergence (T of S) in Longe 10H; time of anthesis (T Anth) and anthocyanin coloration of brace roots (Anth BR) in Longe 7H; and attitude of blade (At of B), anthocyanin coloration of base of glume (Anth Col BG) and anthocyanin coloration of silks (Anth Col Si) in Longe 6H (Table 5). This indicates that there was a degree of phenotypic variation within the hybrids, except Longe 9H. This variation could be attributed to genetic and environmental effects (Idris and Mohammed, 2012). These results are in line with those of Idris and Abuali (2011) who reported significant differences in time to anthesis and ear length for different phonotypic traits in maize genotypes. Different researchers have reported significant amount of variability in different maize populations (Kamara et al. 2003; Grzesiak, 2001 and Khodarahmpour, 2012). These results suggest that the level of variation is higher in Longe 10H followed by Longe 6H and Longe 7H, whereas Longe 9H appeared as a pure variety.

Table 5: Summary statistics of the phenotypic traits that showed significant differences

Abbildung in dieser Leseprobe nicht enthalten

4.1.2 Principal component analysis

The principal component analysis (PCA) revealed four principal components (PCs) for each variety. Syafii et al. (2015) reported four PCs. In Longe 6H, the first four PCs explained a total of 53.3% of the phenotypic variation (Table 6, Appendix 2). The first and the second PCs showed no important traits. In the third PC, which explained 11% of the total variation, the most important traits were angle between main axis and lateral branches (Angle MI) and attitude of lateral branches (At of LB). In the fourth PC, which explained 8% of the total variation, Anthocyanin coloration of silks (An Col of S) and intensity of anthocyanin coloration of silks (In of An Col S) appeared to be important in the fourth PC.

In Longe 7H, the first four PCs explained a total of 51.1% of the phenotypic variation (Table 7, Appendix 3). The first PC revealed no important traits. In the second PC, which explained 19% of the total variation, predominant traits were density of spikelets (De Spi) and number of primary lateral branches (No of PLB). The third PC, which accounted for 13.4% of the total variation, was dominated by traits namely anthocyanin coloration of silks (An Col of S) and ear shape (Ea Sh). The angle between blade and stem (Angle BS), anthocyanin coloration of brace roots (An Col BR), anthocyanin coloration of base of glume (An Col BG), width of blade (W of B) and length of peduncle (Le Peduncle) were important delineating traits associated with the fourth PC, which accounted for 9.9% of the total variation.

In Longe 9H, the first four PCs explained a total of 92.8% of the phenotypic variation (Table 8, Appendix 4). In the second PC, which explained 22.1% of the total variation, predominant traits were the angle between main axis and lateral branches (Angle MI), attitude of lateral branches (At of LB) and time of silking (T of Si). In the third PC, which explained 18.9% of the total variation, Anthocyanin coloration (An Col), anthocyanin coloration of anthesis (An Col of Anth), time of silking (T of Si) and anthocyanin coloration of silks (An Col of S) were the most important traits. While density of spiklets (De Spi) was the only important trait in the fourth PC, which accounted for 11.2% of the total variation.

Abbildung in dieser Leseprobe nicht enthalten

Traits that are corresponding to underlined eigenvalues are the most significant traits that contributed much of the variation in each PC.

Table 7: Longe 7H eigenvalues and cumulative percentage of variation explained by the first four principal components.

Abbildung in dieser Leseprobe nicht enthalten

Traits that are corresponding to underlined eigenvalues are the most significant traits that contributed much of the variation in each PC.

Table 8: Longe 9H eigenvalues and cumulative percentage of variation explained by the first four principal components.

Abbildung in dieser Leseprobe nicht enthalten

Traits that are corresponding to underlined eigenvalues are the most significant traits that contributed much of the variation in each PC.

In Longe 10H, the first four PCs explained a total of 39.1% of the phenotypic variation (Table 9, Appendix 5). In the first PC, which explained 15.1% of the total variation, the most important traits were angle between main axis and lateral branches (Angle MI), time of silking (T of Si) and intensity of anthocyanin coloration of silking (In of An Col S). In the second PC, which explained 8% of the total variation, length of main axis above lowest side branch (Le MALB) and length of main axis above upper side branch (Le MAUB) were the predominant traits. In the third PC, which explained 8.2% of the total variation was dominated by traits such as anthocyanin coloration of glume excluding base (An Col GeB) and anthocyanin coloration of glume of cob (An Col GC). Number of primary lateral branches (No of PLB) and type of grain (Type G) were important in delineating traits associated with the fourth PC, which accounted for 7% of the total variation.

Intensity of anthocyanin coloration of silks (In of An Col S), anthocyanin coloration of base of glume (Anth Col BG), anthocyanin coloration of anthesis (An Col of Anth) and anthocyanin coloration of glume excluding base (Anth Col GeB) had the highest positive loading from Longe 6H, Longe 7H, Longe 9H and Longe 10H, respectively. These traits were important for discriminating the maize hybrids. Chanda et al. (2014) also reported attitude of lateral branches and 6 cm Upper as important traits for discriminating maize inbred lines. The PCA has further proved the existence of the phenotypic variation within maize hybrids.

The interpretation of the results of the PCA is usually carried out by visualization of its PC scores (Zhang et al., 2012). Appendix 2 shows the scores plot of PC1 × PC2 of total samples for Longe 6H. The PC score showed that different samples distributed separately in the two- dimensional area, and agro-dealer samples Muhabura-45, Kigumba-44, Iganga-96, Bweyale43, Masindi-(19, 20, 16, 102, and 15) and Luwero-95 were grouped together with the seed company sample, which represented 40.7% of the total number of the Longe 6H samples which were 27 in number. This implies that 59.3% of the samples were contaminated or not similar to that of the seed company.

Table 9: Longe 10H eigenvalues and cumulative percentage of variation explained by the first four principal components.

Abbildung in dieser Leseprobe nicht enthalten

Traits that are corresponding to underlined eigenvalues are the most significant traits that contributed much of the variation in each PC.

The score plot of Longe 7H (Appendix 3) showed that, agro-dealer samples from Iganga- (106, 105, 112, 107, 114, 27 and 30), Bukedea-42, Soroti-(66, 10 and 2) and Masindi-(105 and 113) were grouped together with the sample from the seed company-A, whereas Iganga- 27 was the only sample grouped together with the seed Company-B. This suggests that there is variation between the seed companies’ samples which might also contribute to the variation within agro-dealers` seeds. This might be due to the fact that, seed companies often contract farmers to grow/multiply seed mostly under rain-fed conditions thereby exposing the crop to variable rainfall conditions (Langyintuo, 2004). Additionally, the grower might face difficulty in achieving isolation distances of 200 - 400m to ensure genetic purity and seed quality (Augustine et al., 2009). Also, the contamination might occur during the processing of the seed. Fifty-eight percent of the total number of the samples (24) were similar to that of the seed company and 42% of the samples were contaminated.

The score plot of Longe 9H (Appendix 4) showed that, only one agro-dealer’ sample (Iganga- 119) was grouped with the seed company’s sample, which represented 33.3% of the total number of Longe 9H samples (6). This implies that the rate of variation from samples got from the seed company was 66.7%. The score plot of Longe 10H showed interesting results (Appendix 5). The seed company’s sample was grouped alone as an outlier. This implies that, the seed company might have made a mistake when packaging the seeds (packaging seeds of another variety in Longe 10H package) which mean the agro-dealers might have got the wrong sample from the seed company, or may be the demand of this variety is high and the quantity produced by the seed company is not enough which forces the agro-dealers to fake the seed.

Agro-dealer samples from Iganga, Masindi, Luwero, Soroti, Bukedea and Muhabura showed high seed quality in comparison with samples from seed companies for the four hybrids under study, whereas samples from Lira, Hoima, Mubende, Mityana, Gulu, Kiboga and Bugiri showed high rate of variation.

4.1.3 Cluster analysis

Results of the cluster analysis (using Euclidean distance and Average link methods) clustered hybrids into three to five main groups. Longe 10H samples grouped into four groups with similarity level of 85%. The first cluster included only the seed company’s sample (reference). The second cluster contained 47 agro-dealer samples, which were sub-grouped into three different groups. The third group contained one agro-dealer sample, while the fourth group contained two agro-dealer samples (Figure 1). Longe 9H clustered into four groups with 80% level of similarity, the first group contained two agro-dealer samples, including the seed company’s sample. The second group contained two agro-dealer samples, the third group contained one agro-dealer sample as well as the fourth group (Figure 2). Longe 6H clustered into sex main groups with 85% similarity level. The first group contained 19 agro-dealer samples, which included the seed company’s sample. The second group contained one agro-dealer sample as well as the fifth group. The third group contained 2 agro- dealer samples, while the fourth group contained four agro-dealer samples (Figure 3). The results indicated that there was variation within each variety based on the phenotypic data. Longe 10H was the variety with the highest level of variation (98%) based on the cluster analysis, followed by Longe 7H (75%), Longe 6H (70%) and Longe 9H (67%). This might be because of the high demand for Longe 10H and the seed company cannot supply the required amount.

Cluster analysis of Longe7H showed the most interesting results (Figure 4). Longe 7H clustered into four groups with 85% level of similarity. The first group contained 17 agro- dealer samples, including that of seed Company-B. The second group contained one agro- dealer sample as well as the fourth group. The third group contained four agro-dealer samples and that of the seed company-A (Figure 4). Samples of Longe 7H which is produced commercially by two seed companies were expected to be clustered together. Surprisingly, they clustered into two different groups. This suggests that the seed companies themselves might be the source of variation by selling different versions of seed to the agro-dealers. This might be due the agro-dealer or the seed company agents’ problems such as lack of credibility on the part of some agents (adulteration of seed) and poor storage facilities.

Abbildung in dieser Leseprobe nicht enthalten

In Longe 6H, the seed company’s sample clustered into sub-group with Masindi-99, 24 and Iganga-24. Agro-dealers in these districts were close to the seed company in terms of geographic location. In Longe 7H, the seed company-A clustered into sub-group with Mityana-116, Iganga-26,106 and 115 and Lurero-111, and seed Company-B cluttered with Masindi-113, which are also closer to the seed companies. This suggests that, the further the agro-dealers are from the source (seed company), the more contaminated is their seed. This could be due to the fact that, agro-dealers face problems such as poor road networks, limited access to transportation facilities and poorly established storage infrastructure in the rural areas. Some agro-dealers are known to willfully adulterate the seed by opening the seed packs, taking out portions and mixing the rest with grain, which consequently compromises the quality of the seed. Furthermore, agro-dealers often have poor storage facilities. In fact most of them store seed, fertilizer and other agro-inputs side by side in the same storage facilities for extended periods thereby compromising the quality of seeds (Augustine et al., 2009). Cluster analysis of Longe 10H agreed with the PCA score plot of the same variety.

The study also estimated the percentage of off-types in each variety (the number of off-types ×100, then divided by the total number of sampled plants). The results proved that, agro- dealers might mix the improved seed with grain in order to increase the seed quantity and make more profit (Table 10). Over staying in the storage of the improved seed might be another reason for the variation found at agro-dealers level, which might explain the germination failure of some samples in the field (loss of seed vigour) as reported by Barnett et al. (2011).

Table 10: Percentage of off-types within the field

Abbildung in dieser Leseprobe nicht enthalten

4.2 Genetic variation

Of the 142 SNP’s used for genotyping, 128 SNP’s were successfully genotyped and able to show good amplification and reproducible in allele calls, hence included in the data analysis.

These 128 SNPs detected a total of 256 alleles, with each SNP detecting two alleles as expected. Appendix 6 gives the summary statistics of 120 SNPs (including number of markers, major allele frequency and gene diversity).

4.2.1 Allele frequency

The major allele frequency ranged from 0.509 to 1.000, Santacruz-varela, (2001) reported major allele frequency of 1.000 in North American maize. In Longe 10H, 46 samples were successfully genotyped out of a total number of 63 samples. In Longe 9H, out of a total number of 6 samples, 4 samples were successfully genotyped, while 11 samples were successfully genotyped out of a total number of 24 samples in Longe 7H. For Longe 6H, the seed company’s sample was not genotyped, and as a result, Longe 6H samples were not included in the analysis.

4.2.2 Heterozygosity

Heterozygosity gives an idea of the information available from the SNPs loci and their potential to detect differences within/among hybrids based on their genetic relation. The study clearly showed the presence of high genetic heterogeneity within the hybrids under investigation. In the entire collection, the proportion of heterogeneity varied from 0.00 to 95.5 % and the average was 43.3 % (Appendix 7). Longe 10H samples had the highest heterogeneity level (46 %) among the hybrids under study, which is three-fold the considerable heterogeneity level in hybrid (≤ 15% heterogeneity). Whereas, Longe 7H demonstrated 2.7-fold the expected, Longe 9H demonstrated 2.5-fold the expected, and Longe 6H with 35 % heterogeneity, as the lowest showed 2.3-fold the expected. However, the heterozygosity of all hybrids under study almost the same. Gethi et al. (2002) reported heterogeneity value 5-fold in B73 from Cornell University and 2.5 fold the expected value in maize. Semagn et al. (2012) observed that, contamination was the most likely cause for the observed differences in heterogeneity between the seed sources of the hybrids. The cause might also be due to the inbred lines of these hybrids. When the same inbred line is grown in different environments, some of the loci that have undergone genetic changes may contain alleles that are latent in one environment but are expressed in another environment (Gethi et al., 2002). This phenomenon is referred to as ascertainment bias in population genetics. It has been reported that missing calls may lead to biased conclusions in the estimation of allele frequencies, and therefore in the estimation of heterogeneity percentage (Fu et al., 2009). However, these results are in agreement with the phenotyping results of Longe 10H, and therefore confirm that Longe 10H has the highest level of genotypic and phenotypic variation from the seed company’s sample.

4.2.3 Genetic distance

Table 11 presents codes of samples with their districts of origin as appearing in the genetic distance matrix. In the present study, genetic distance between seed sources of Longe 7H varied from 0.026 to 0.273 (Table 12). Samples L7H-26 obtained the greatest distance from the seed company sample (L7H-47-SC-A) which was 0.273, while the shortest distance was 0.032 obtained by L7H-2. The greatest distance from L7H-50-SC-B (sample from seed Company-B) was 0.179 obtained by L7H-47-SC-A. This indicates that, there is high genetic variation between samples from seed company-A and seed Company-B. The shortest distance from L7H-50-SC-B was 0.081 obtained by L7H-107. The agro-dealer sample L7H-26 which scored the highest distance from our seed company-A, scored 0.142 distance from the seed Company-B. This implies that, the agro-dealer samples of Longe7H had less genetic variation from the seed Company-B than from the seed company-A, which supported the phenotyping results (Figure 4). The only reason could be that; the two seed companies are selling different versions of the same variety as was explained earlier.

Table 11: Codes of samples with their districts of origin as appearing in the genetic distance matrix

Abbildung in dieser Leseprobe nicht enthalten

Table 12: Genetic similarity matrix of Longe 7H based on genetic distance

Abbildung in dieser Leseprobe nicht enthalten

Longe 9H varied from 0.030 to 0.254, with 0.214 as the greatest distance from the seed company’s sample (L9H-49-SC) which was obtained by L9H-120 while the closest distance was 0.029, which was obtained by L9H-1 (Table 13). Genetic distance for Longe 10H varied from 0.004 to 0.334. The closest distance to the seed company sample L10H-51-SC was 0.044 obtained by L10H-9 and the greatest distance was 0.263 obtained by L10H-74 (Table 14). Dao et al. (2014) investigated the extent of genetic differentiation among a set of 96 maize inbred lines, using 1057 SNP markers, and reported genetic distance of 0.314. Lanes et al., (2014) reported genetic distance of 0.62 which was a little higher than the findings of this study. These results imply that, there are significant differences between the agro-dealers’ samples which was also explained by Igniatovic-Micic et al., (2003) who compared RFLP and RAPD techniques for effectiveness in maize population characterization and identification. They also reported genetic distance range from 0.131 to 0.456.

The greatest genetic distances recorded were 0.273, 0.263 and 0.214 in Longe 7H, Longe 10H and Longe 9H, respectively. This suggests that, the amount of variation is higher between Longe 7H samples followed by Longe 10H and Longe 9H has the lowest level of variation.

Table 13: Genetic similarity matrix of Longe 9H based on genetic distance

Abbildung in dieser Leseprobe nicht enthalten

Based on the genetic distance, agro-dealer’s samples Soroti (L9H-1, L10H-9 and L7H-2) showed the smallest genetic distance in Longe 9H, Longe 10H and Longe 7H, respectively. That might imply, that the agro-dealers from Soroti have a good control on their seed marketing system, plus other social factors. Thus, their seed has the highest genetic purity when compared with samples from other agro-dealers. In contrast, samples got from agro- dealers in Luwero district showed high genetic variation/contamination among the agro- dealers.

Table 14: Genetic similarity matrix of Longe 10H based on genetic distance

Abbildung in dieser Leseprobe nicht enthalten

4.2.4 Cluster analysis

In order to gain further insight into the genetic diversity within the different groups of maize hybrids, a UPGMA tree based on Roger’s genetic distance was constructed (Tamura et al., 2011). Cluster analyses showed clear genetic variation within each hybrid. Longe 7H clustered into two main groups. Group one which included 81.8% of the total number of the samples clustered into more 6 sub-groups (Figure 7). The sample from seed company-A (L7H-47) clustered with the sub-group that represented 63.6% of the total number of the samples, whereas the sample from seed Company-B (L7H-50) clustered with the sub-group that represented 18.9% of the total number of samples (Figure 5). Longe 9H clustered into two different groups, with 75% of the agro-dealer samples clustering with the seed company sample in group one (Figure 6). Group one also clustered into another two different groups, with 50% of the agro-dealer samples similar to that of the seed company (Figure 6). These results are in agreement with the phenotypic analysis which allows us to conclude that the agro-dealers’ samples were highly similar to that of the seed Company-A.

The cluster analysis of Longe 10H is completely agreed with the phenotypic cluster of the same hybrid. In which clustered Longe 10H into three major groups, with 90.7% of the samples in group one including the seed company’s sample L10H-51 (reference sample) (Figure 7). However, group one sub-clustered into more than 10 groups, with only 4.4% (which represented 2 samples) of the agro-dealer samples grouped with the reference sample (Figure 5). It can be seen that samples from group one is quite similar, but with slight variation (10 sub-groups). This might be because, the agro-dealers could be getting seed from different multiplication years or different lot numbers. For example, the agro-dealer might get a big amount of seed and some of it might remain and be sold to farmers in the next planting season.

Figure 5: Longe 7H dendrogram, based on the genetic distance using Roger’s similarity matrix and UPGMA clustering method.

Figure 5: Longe 9H dendrogram, based on the genetic distance using Roger’s similarity matrix and UPGMA clustering method.

Figure 7: Longe 10H dendrogram, based on the genetic distance using Roger’s similarity matrix and UPGMA clustering method.

In comparison with phenotypic clustering, the average link cluster of Longe 9H perfectly agreed with the clustering based on the genetic distance. However, the phenotypic clustering of Longe 7H was different from the genetic clustering, but still reflecting the same finding which is the existence of variation. This could be explained by the fact that, in cluster analysis, different combinations of genetic distance/similarity matrix and clustering algorithms can give rise to different groups (Dao et al., 2014). In addition, the clustering agreed with the genetic distance as Soroti (L9H-1, L7H-2) clustered with the reference (L9H- 49, L7H-47) for the two hybrids. Moreover, Iganga’s samples (L9H-119, L7H-107 and L10H- 76) also clustered with the reference samples (L9H-49, L7H-47 and L10H-51) in the three hybrids under study. This could be because of the closer geographic distance to the sources, as was explained above or credibility of agro-dealers. Luwero’s samples continued to appear as the most contaminated. Despite, the small sample size that was used in this study, the results clearly show the existence of variation at agro-dealer level, but in order to explain this variation further studies are necessary.

CHAPTER FIVE CONCLUSIONS AND RECOMMENDATIONS

In this study, the presence of variation in the agro-dealer seeds of four common maize hybrids (Longe 6H, 7H, 9H and 10H) in Uganda seed market was investigated using phenotypic and genotypic characterization methods. A total of 120 samples (116 samples from agro-dealers in 15 districts and 4 samples from seed companies) were used in this study. The seeds were sown in Alfa lattice design. Phenotypic data were collected from 10 randomly selected plants for each sample using 33 descriptors developed by CIMMYT. Genotyping was carried out on DNA extracted from leaves of 10-14 day old seedlings and 128 SNPs were used. The analysis of variance (ANOVA), principle component analysis and cluster analysis based on the phenotypic traits and the genetic traits were used to explore whether the agro-dealer seeds were pure or contaminated.

The results of ANOVA revealed significant differences (P≤ 0.05) in 11 phenotypic traits out of 33 traits that were studied. These were time of anthesis (T Anth), density of spiklets (De Spi), length of main axis above lowest side branch (Le MALB), ear length (E Le), length of peduncle (Le of Ped) and time of silk emergence (T of S) in Longe 10H; time of anthesis (T Anth) and anthocyanin coloration of brace roots (Anth BR) in Longe 7H; and attitude of blade (At of B), anthocyanin coloration of base of glume (Anth Col BG) and anthocyanin coloration of silks (Anth Col Si) in Longe 6H.

The PCA showed that, Longe 10H has the highest rate of variation (95%) in the agro-dealers’ samples from the seed companies sample followed by Longe 9H (66.7%), Longe 6H (59.3%) and Longe 7H (41.7%). Intensity of anthocyanin coloration of silks (In of An Col S), attitude of lateral branches (At of LB) and the angle between blade and stem (Angle BS) in Longe 6H; width of blade (W of B), anthocyanin colouration of silks (An Col of S) and density of spikelets (De of Spi) in Longe 7H; anthocyanin colouration of anthesis (An Col Anth), time to silk emergence (T of Si) and anthocyanin colouration of glumes of cob (Anth Col GC) in Longe 9H; and anthocyanin colouration of glumes of cob (Anth Col GC), anthocyanin colouration of glumes excluding base (An Col Gex), and time of silk emergence (T of Si) in Longe 10H had the highest positive loading. These traits were important for discriminating the maize hybrids.

The genetic distance analysis showed that agro-dealers’ samples from Soroti and Iganga had the highest seed quality (genetically pure and similar to the seed companies’ seed) for all the hybrids under the study. Agro-dealers’ sample from Luwero had the highest rate of contamination among the agro-dealer`s samples. Longe 10H had the highest level of variation with 46% heterogeneity followed by Longe 7H (41.5%), Longe 9H (38%) and Longe 6H (35.4%). All these different findings supported the presence of variation/contamination within the hybrids under study at agro-dealer level.

Based on the results the following conclusions and recommendations were made:

Agro-dealers’ samples from Iganga, Masindi, Soroti, Bukedea and Muhabura showed high seed quality level in comparison with samples from the seed companies for the four hybrids under study, whereas samples from Lira, Hoima, Mubende, Mityana, Gulu, Kiboga, Bugiri and Luwero showed high rate of variation. Therefore, further study is required to investigate the agro-dealers in these districts. The investigation can be carried out by occasionally sending seed inspectors either from the Seed Certifying Agency (for official inspection) or from the Seed Company (internal or accredited inspection) to these areas to collect seed samples from the agro-dealers for evaluation.

Seed companies have to work with trust worthy dealers otherwise companies will lose customers. To avoid long time storage of seed or selling different versions of seed, seed companies could put the year and season of production on the packages so that farmers can differentiate the seeds.

The phenotypic and genotypic cluster analysis of Longe 7H clearly showed that, variation might occur before the seed leaves the seed company. Therefore, the seed quality control and certification department need to be vigilant in monitoring the processing of seed.

REFERENCES

Agona, A. and Muyinza, J.N.H. (2001). An overview of maize in Uganda. Postharvest Programme. Retrieved from http://www.mendeley.com/research/overview-maize- uganda/

Ahmed, M. (2012). Analysis of incentives and disincentives for maize in Uganda, december 2012. Technical Notes Series, MAFAP, FAO, (December). http://doi.org/Technical notes series, MAFAP, FAO,

Akaogu, I.C., Adetimirin, V.O. and Oyekunle, M. (2013). Genetic diversity assessment of extra-early maturing yellow maize inbreds and hybrid performance in Striga -infested and Striga -free environments. Journal of Agricultural Science, 151: 519-537. http://doi.org/10.1017/S0021859612000652

Amankwaah, V.A. (2012). Phenotypic and molecular characterization of released and elite sweet potato varieties in Ghana compared with virus-tested putative ramets. University of science and technology. Retrieved from http://www.mendeley.com/research/phenotypic-molecular-characterization-released- elite-sweet-potato-varieties-ghana-compared-virustest/

Augustine, L., Abdoulaye, T., Samogo, D., Simeon., Bamire, A. and Adetuiji, O. (2009) Assessing the contrants affecting production and deployment of maize seed in DTMA countries of west Africa. IITA, Ibadan, Nigeria. 40 pp.

Babu-Apraku, B., Fakorede, M.A.B., Menkir, A., Sanogo, D. (2012). Conduct and management of maize field trials. Citeseer, Pp 59.

Bamaiyi, J., Oniemayin, M., (2011). Management of stem borers on some quality protein maize varieties. Journal of Agricultural Sciences, 56(3): 197-205.

Bänziger, M. (2001). Maize research highlights 1999-2000. CIMMYT. Mexico City. Retrieved from http://www.mendeley.com/research/maize-research-highlights- 19992000-cimmyt-mexico-city/

Barnett, C., Chisvo, M. and Pinto, Y. (2011). Country case studies on the pass value chain strategy / approach and its impact / effect on smallholder farmer Prepared by : EAST AFRICA SYNTHESIS REPORT, (January).

Benson, J.M., Poland, J.A., Benson, B.M., Stromberg, E.L., Nelson, R.J. (2015). Resistance to Gray Leaf Spot of Maize: Genetic Architecture and Mechanisms Elucidated through Nested Association Mapping and Near-Isogenic Line Analysis. PLoS Genet 11(3): e1005045. doi:10.1371/journal.pgen.1005045.

Botstein, D., White, R.L., Skolnick, M. and Davis, R.W. (1980). Construction of a genetic linkage map in man using restriction fragment length polymorphisms. American Journal of Human Genetics, 32 (3): 314-31. Retrieved from http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1686077&tool=pmcentrez &rendertype=abstract Bucheyeki, T.L. (2012). Characterization and genetic analysis of maize germplasm for resistance to northern corn leaf blight disease in Tanzania. Thesis, University of KwaZulu-Natal, Pietermaritzburg, South Africa. Unpublished.

Bucheyeki, T.L., Gwanama, C., Mgonja, M., Chisi, M. and Folkertsma, R. (2008).

Morphological Characterization of Tanzanian Sorghum [ Sorghum bicolor (L) Moench] Landraces. Kasetsart J. (Natural Science.), 588: 579-588.

Buckler, E.S., Ganal, M.W., Durstewitz, G., Polley, A., Clarke, J.D., Graner, E. and Falque, M. (2011). A Large Maize (Zea mays L.) SNP Genotyping Array : Development and Germplasm Genotyping , and Genetic Mapping to Compare with the B73 Reference Genome. PLoS ONE, 6 (12): e28334 http://doi.org/10.1371/journal.pone.0028334

Cadee, N. (2000). Genetic and environmental effects on morphology and fluctuating asymmetry in nestling barn swallows. Journal of Evolutionary Biology, 13: 359-370. http://doi.org/10.1046/j.1420-9101.2000.00194.x

Chanda, R., Mukanga, M., Mwala, M., Osiru, D.S. and Macrobert, J. (2014). A comparative analysis of distinctness , uniformity and stability ( DUS ) data in discriminating selected Southern African maize (Zea mays L.) inbred lines. African Journal of Agreculrural Research, 9 (41): 3056-3076. http://doi.org/10.5897/AJAR2014.8755

Charcosset, A. and Moreau, L. (2004). Use of molecular markers for the development of new cultivars and the evaluation of genetic diversity. Euphytica, 137: 81-94.

Dao, A. (2013). Genetic characterization, association mapping and evaluation of heterotic patterns of tropical maize (Zea mays L.) Inbred lines under drought and non-drought conditions, University of Ghana. Ghana. Unpublished.

Dao, A., Sanou, J., Mitchell, S.E., Gracen, V. and Danquah, E.Y. (2014). Genetic diversity among INERA maize inbred lines with single nucleotide polymorphism (SNP) markers and their relationship with CIMMYT, IITA, and temperate lines. BMC Genetics, 15 (1): 127-141. http://doi.org/10.1186/s12863-014-0127-2

Dekkers, J.C.M. and Hospital, F. (2002). The use of molecular genetics in the improvement of agricultural populations. Nat Rev Genet, 3 (1): 22-32.

Devi, H.N., Devi, K.N., Singh, N.B., Singh, T.R., Jyotsna, N. and Paul, A. (2013). Phenotypic Characterization , Genetic Variability and Correlation Studies among Maize Landraces of Manipur. International Journal of Bio-Resource and Stress Management, 4 (2): 352-355.

Devos, K.M. and Gale, M.D. (2000). Genome relationships: the grass model in current research. The Plant Cell, 12 (5): 637-646. http://doi.org/10.1105/tpc.12.5.637

Elepu, G. and Dalipagic, I. (2014). Agricultural value chain analysis in northern uganda : Action Against Hunger | ACF-International, (March). Retrieved from http://www.mendeley.com/research/agricultural-value-chain-analysis-northern- uganda-maize-rice-groundnut-siunflower-sesame/

FAOSTAT (2010). Statistical databases and data-sets of the Food and Agriculture Organization of the United Nations.

FAOSTAT (2012). Food and agriculture organization of the united nations. Retrieved from http://faostat.fao.org/site/339/default.aspx

Fetahu, S., Kaçiu, S., Aliu, S., Bajraktari, I., Rusinovci, D.Z.I. and Salihu, S. (2012). Genetic and Phenotypic Diversity among Some Common Bean Landraces (Phaseulus vulgaris L.) in Kosovo. Acta Horticulturae, 1: 169-174.

FNSU-UDS. (2011). Three Country (Burkina Faso, Ghana and Mali) Case Studies on the PASS Value Chain Strategy / Approach and its Effect on Smallholder Farmers Yields in Africa. Draft report submitted to Alliance for a Green Revolution in Africa (AGRA) Food and Nutrition. Retrieved from http://www.mendeley.com/research/three-country- burkina-faso-ghana-mali-case-studies-pass-value-chain-strategy-approach-effect- smallho/

Fraley, R.T. (2009). Molecular Genetic Approaches to Maize Improvement - an Introduction. Biotechnology in Agriculture and Forestry, 63: 3-7.

Fu, W., Wang, Y., Wang, Y., Li, R., Lin, R. and Jin, L. (2009). Missing call bias in high- throughput genotyping. BMC Genomics, 10 (1): 106. http://doi.org/10.1186/1471- 2164-10-106

Gabriel, L.C., Maximo, B.L. and Ruggero, B. (2009). Heterosis and heterotic patterns among maize landraces for forage. Crop Breeding and Applied Biotechnology: 229-238.

GenStat 14.1. 2012. GenStat 14th edition for Windows. VSN Int., Oxford, UK. https://www.vsni.co.uk/downloads/genstat/14th-edition-upgrade/

Gethi, J., Gethi, J.G., Labate, J.A., Lamkey, K.R., Smith, M.E. and Kresovich, S. (2002). SSR Variation in Important U . S . Maize Inbred Lines. Crop Science, 42: 951-957. http://doi.org/10.2135/cropsci2002.0951

Gichuru, L.N. (2013). Breeding investigations on utility of maize streak virus. PhD Thesis, University of KwaZulu-Natal, (November). Unpublished.

Grattapaglia, D., Silva-Junior, O.B., Kirst, M., de Lima, B.M., Faria, D.A. and Pappas, G.J. (2011). High-throughput SNP genotyping in the highly heterozygous genome of Eucalyptus: assay success, polymorphism and transferability across species. BMC Plant Biology, 11(1): 65-83. http://doi.org/10.1186/1471-2229-11-65

Grzesiak, S. (2001). Genotypic variation between maize (Zea mays L.) single cross hybrids in response to drought stress. AICA Physiologiae Plantarum, 23 (4): 443-456.

Hamblin, M.T., Warburton, M.L. and Buckler, E.S. (2007). Empirical Comparison of Simple Sequence Repeats and Single Nucleotide Polymorphisms in Assessment of Maize Diversity and Relatedness. PLoS ONE, 2 (12): e1367. http://doi.org/10.1371/journal.pone.0001367.

Hoisington, D., Khairallah, M. and Gonzalez-de-leon, D. (1998). Laboratory protocols.

CIMMYT Applied Molecular Genetics Laboratory, Mexico.

Idris, A.E. and Abuali, A.I. (2011). Genetic variability for vegetative and yield traits in maize (Zea mays L.) genotypes. International Research Journal of Agricultural Science and Soil Science, 1:408-411.

Idris, A.E. and Mohammed, H.I. (2012). Screening Maize (Zea mays L.) Genotypes by Genetic Variability of Vegetative and Yield Traits Using Compromise Programming Technique. British Biotechnology Journal, 2 (2): 102-114.

Igniatovic-Micic, D., Coric, T., Kovacavic, D., Markovic, K. and Lazic-Jancic, V. (2003). RELP and RAPD analysis of maize (zea mays L.) local populations for identification of variablity and duplicate accessions. Maydica, 48: 153-159.

Ignjatović-Micić, D., Ristić, D., Babić, V., Andjelković, V., Marković, K. and Vančetović, J. (2013). Genetic assessment of maize landraces from former Yugoslavia. Genetika, 45 (2): 405-417. http://doi.org/10.2298/GENSR1302405I

Ribaut, J.M., Hoisington, D., Ba¨nziger, M., Tim L. and Setter, G.O.E. (2004). Genetic Dissection of Drought Tolerance in Maize: A Case Study. Physiology and Biotechnology Integration for Plant Breeding, 15: 571-609. Retrieved from http://www.dekker.com

Kamara, A.Y., Menkir, A., Badu-Apraku, B. and Ibikunle, O. (2003). Reproductive and stay- green trait responses of maize hybrids, improved open-pollinated varieties and farmers’ local varieties to terminal drought stress. Maydica, 48 (1): 29-37.

KASPar protocol http://www.ksre.ksu.edu/igenomics/doc1363.ashx. (n.d.).

Khodarahmpour, Z. (2012). Morphological Classification of Maize (Zea mays L.) Genotypes in Heat Stress Condition. Journal of Agricultural Science, 4 (5): 31-40. http://doi.org/10.5539/jas.v4n5p31

Khoza, S. (2012). Assessment of Maize Germplasm Lines for Genetic Diversity, Cultivar Superiority and Combining Ability. Doctoral Dissertation, University of KwaZulu- Natal, Pietermaritzburg, (November). Unpublished.

Kumar, M., Fougat, R.S., Sharma, A.K., Kulkarni, K., Mistry, J.G., Sakure, A.A. & Kumar, S. (2014). Phenotypic and molecular characterization of selected species of Plantago with emphasis on Plantago ovata. Australian Journal of Crop Science, AJCS, 8 (12): 1639- 1647.

Lafarge, T.A. and Hammer, G.L. (2002). Tillering in grain sorghum over a wide range of population densities: Modelling dynamics of tiller fertility. Annals of Botany, 90 (1989): 99-110. http://doi.org/10.1093/aob/mcf153

Lanes, E.C.M., Viana, J.M.S., Paes, G.P., Paula, M.F.B. and Maia, C. (2014). Population structure and genetic diversity of maize inbreds derived from tropical hybrids. Genetics and Molecular Research, GMR, 13 (3): 7365-7376.

Langyintuo, A.S. (2004). Challenges of the Seed Sector of Southern Africa. International Maize and Wheat Improvement Center (CIMMYT) Report, Harare, Zimbabwe, 39. Retrieved from <www.cimmyt.org/gis/rfseedsafrica/>

Langyintuo, A.S., Mwangi, W., Diallo, A.O., MacRobert, J., Dixon, J. and Bänziger, M. (2009). An analysis of the challenges of the maize seed industry in eastern and southern Africa. Journal of Chemical Information and Modeling, 53: 1-36. http://doi.org/10.1017/CBO9781107415324.004

Larson, D.W. and Mbowa, S. (2004). Strategic marketing problems in the Uganda maize seed industry. International Food and Agribusiness Management Review, 7 (4): 86-93.

Liu, K. and Muse, S.V. (2005). Power Marker: an integrated analysis environment for genetic marker analysis. Bioinformatics, 21 (9): 2128-2129. http://doi.org/10.1093/bioinformatics/bti282

Magorokosho, C. (2006). Genetic Diversity and Performance of Maize Varieties From Zimbabwe, Zambia and Malawi. PhD Dissertation, Texas A&M University, (December), 194pp. Unpublished.

Mohamed, N.S. (2013). Insect damage: Damage on Post-harvest. International Centre of Insect Physiology and Ecology (ICIPE), INPhO (www.icipe.org).

Mohammadi, S. and Prasanna, B.M. (2003). Analysis of genetic diversity in crop plants- salient statistical tools and consideration. Crop Science, 43(4): 1235-1248.

Mubangzi, E., Ntamu, D.N., Thembo, Wilfred, M. and Thijssen, M. (2012). ISSD Briefing Note - September 2012 Uganda Seed Sector Assessment. ISSD Africa, (September), p. 1-7. Murray, M.G., and Thompson W.F. (1980). Rapid isolation of high molecular weight plant DNA. Nucleic Acids Research. 8:4321-4325.

Ndhlela, T. (2012). Improvement strategies for yield potential, disease resistance and drought tolerance of Zimbabwean maize inbred lines. PhD Thesis, University of the Free State, South Africa, (June). Unpublished.

Nyombayire, A., Edema, R., Asea, G. and Gibson, P. (2011). Combining ability of maize inbred lines for performance under low nitrogen and drought stresses. African Crop Science Conference Proceedings, 10: 579-585.

PELUM, M. (2012). The seed sector of Uganda. Pelum Assciation.

Parida, S.K., Mukerji, M., Singh, A.K., Singh, N.K. and Mohapatra, T. (2012). SNPs in stress-responsive rice genes: validation, genotyping, functional relevance and population structure. BMC Genomics, 13 (1): 426-443. http://doi.org/10.1186/1471- 2164-13-426

Prasanna, B.M. and Hoisington, D. (2003). Molecular breeding for maize improvement: An overview. Indian Journal of Biotechnology, 2 (1): 85-98.

Prasanna, B.M., Pixley, K., Warburton, M.L. and Xie, C.X. (2010). Molecular marker- assisted breeding options for maize improvement in Asia. Molecular Breeding, 26 (2): 339-356. http://doi.org/10.1007/s11032-009-9387-3

Ribaut, J., Gregory, O., Edmeades, G.O., Hoisington, D., Tim, L., Setter, and Banziger, M. (2004). Genetic dissection of drought tolerance in maize. Physiology and Biotechnology Integration for Plant Breeding. 571 - 609 pp.

Rijsberman, F. R. (2006). Water scarcity: fact of fiction? Agricultural Water Management, 80(1 -3): 2 -22.

Rogers, J.S. (1972). Measures of genetic similarity and genetic distance. Studies in Genetics, 7 (7213): 145-153.

Ruizd, G. and Alvarez, A. (2001). Morphological classification of maize landraces from Northern Spain. Genetic Resources and Crop Evolution, 48: 391-400. http://doi.org/10.1023/A:1012074606561

Salgado, K. (2006). Genetic purity certificate in seeds of hybrid maize using molecular, Revista Brasileira de Sementes, 28: 169-175.

Santacruz-varela, A. (2001). Genetic diversity of North American popcorn and its relationship with Mexican and South American popcorns - Dissertations & Theses @ Iowa State University - ProQuest. Retrospective Theses and Dissertations., (1078). Retrieved from http://search.proquest.com.proxy.lib.iastate.edu/pqdtlocal1005976/docview/27589043 1/DF68BFF6EFA4F2BPQ/15?accountid=10906

Semagn, K., Beyene, Y., Makumbi, D., Mugo, S., Prasanna, B.M., Magorokosho, C. and Atlin, G. (2012). Quality control genotyping for assessment of genetic identity and purity in diverse tropical maize inbred lines. TAG. Theoretical and Applied Genetics.

Theoretische Und Angewandte Genetik, 125 (7): 1487-501. http://doi.org/10.1007/s00122-012-1928-1

Semagn, K., Magorokosho, C., Ogugo, V., Makumbi, D. and Warburton, M.L. (2014). Genetic relationships and structure among open-pollinated maize varieties adapted to eastern and southern Africa using microsatellite markers. Molecular Breeding, (34): 1423-1435. http://doi.org/10.1007/s11032-014-0126-z

Shiferaw, B., Prasanna, B.M., Hellin, J. and Bänziger, M. (2011). Crops that feed the world 6. Past successes and future challenges to the role played by maize in global food security. Food Security, 3: 307-327. http://doi.org/10.1007/s12571-011-0140-5

Ssebuliba, R.N. (2010). Baseline survy of the seed sector in Uganda in relation to regional harmonization of seed legislation. Uganda Seed Trade Association (USTA).

Statista. (2013). The Statistics Portal-Global Corn Production. Retrieved from http://www.statista.com/statistics/254292/global-corn-production-by-country/

Syafii, M., Cartika, I. and Ruswandi, D. (2015). Multivariate Analysis of Genetic Diversity among some Maize Genotypes under Maize-Albizia Cropping System in Indonesia.

Asian Journal of Crop Science, 7 (4): 244-255. http://doi.org/10.3923/ajcs.2015.244.255

Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. and Kumar, S. (2011). MEGA5: Molecular evolutionary genetics analysis using maximum likelihood, evolutioanry distance, and maximum parsimony methods. Molecular Biology and Evolution, 28 (10): 2731-2739. http://doi.org/DOI 10.1093/molbev/msr121

Tenaillon, M.I., Sawkins, M.C., Long, A.D., Gaut, R.L., Doebley, J.F. and Gaut, B.S. (2001). Patterns of DNA sequence polymorphism along chromosome 1 of maize (Zea mays L.). Proceedings of the National Academy of Sciences of the United States of America PNAS, 98 (16): 9161-9166. http://doi.org/10.1073/pnas.151244298

Uganda AGRA-PASS. (2010). Mid-Term Review of the Program for Africa’s Seed Systems UGANDA Submitted to the The Alliance for a Green Revolution in Africa. Nairobi.

Van Inghelandt, D., Melchinger, A.E., Lebreton, C. and Stich, B. (2010). Population structure and genetic diversity in a commercial maize breeding program assessed with SSR and SNP markers. Theoretical and Applied Genetics, 120: 1289-1299. http://doi.org/10.1007/s00122-009-1256-2

Verheye, W. (2010). Growth and Production of Maize: Traditional Low-Input Cultivation. Encyclopedia of Life Support Systems (EOLSS), 2: p.1.

Vicente, M.C. De, Guzmán, F.A, Engels, J. and Rao, V.R. (2005). Genetic Characterization and Its Use in Decision Making for the Conservation of Crop Germplasm. Management, p. 121-128.

Vogel, K.E. (2010). Genetic characterization of maize populations selected for grain methionine content. Lowa State Universit y, (Thesis). Unpublished.

Weishi, C., Chunqing, Z. and Jinxing, W. (2012). Purity identification of maize seed based on discrete wavelet transform and BP neural network. Transactions of the Chines Socity of Agricultural Engineering, 28 (2): 253-258.

Yan, J., Shah, T., Warburton, M.L., Buckler, E.S., Mcmullen, M.D. and Crouch, J. (2009). Genetic Characterization and Linkage Disequilibrium Estimation of a Global Maize Collection Using SNP Markers. PLoS ONE, 4 (12): e8451. http://doi.org/10.1371/journal.pone.0008451

Zhang, X., Liu, F., He, Y. and Li, X. (2012). Application of Hyperspectral Imaging and Chemometric Calibrations for Variety Discrimination of Maize Seeds. Sensors, 12 (12): 17234-17246. http://doi.org/10.3390/s121217234.

LIST OF APPENDICES

Appendix 1: Summary statistics of the phenotypic traits, MS, Average SED and Average LSD

Abbildung in dieser Leseprobe nicht enthalten

*M.S= Mean of squires, SED = standard error of difference, LSD= least significant difference, *=Significant at P ≤ 0.05,

**=Significant at P ≤ 0.01, Significant at P ≤ 0.001.

Appendix 2: Principal Component Analysis (PCA) score plot for Longe 6H based on the 33 phenotypic traits

Abbildung in dieser Leseprobe nicht enthalten

Appendix 3: Principal Component Analysis (PCA) score plot for Longe 7H based on the 33 phenotypic traits

Abbildung in dieser Leseprobe nicht enthalten

Appendix 4: Principal Component Analysis (PCA) score plot for Longe 9H based on the 33 phenotypic traits

Abbildung in dieser Leseprobe nicht enthalten

Appendix 5: Principal Component Analysis (PCA) score plot for Longe 10H based on the 33 phenotypic traits

Abbildung in dieser Leseprobe nicht enthalten

Appendix 6: Summary statistics of 120 SNPs.

Abbildung in dieser Leseprobe nicht enthalten

Appendix 7: Summary of Heterozygosity (%) and PIC (%) of Longe 6H, 7H, 9H and 10H

Abbildung in dieser Leseprobe nicht enthalten

PIC= Polymorphism Information Content

¹

Figure 1: Longe 10H Dendrogram, based on 33 phenotypic traits using Euclidean distance matrix and Average link method

¹ Agro-dealers are represented by districts S.C: Seed company sample (reference)

² Agro-dealers are represented by districts S.C: Seed company sample (reference)

³ Agro-dealers are represented by districts S.C: Seed company sample (reference)

⁴ Agro-dealers are represented by districts S.C: Seed company sample (reference)

⁵ SC-A= seed company-A

⁶ SC-B= seed company-B

⁷ Seed company’s sample

⁸ Seed company’s sample Table 14 (continued)