Chapter 3 Genetic resources of cassava: potential of breeding for improving storage potential

Contents - Previous - Next

Assembling and characterizing existing genetic diversity
Genetics and cytogenetics
Creation of new genetic diversity
Objectives in cassava breeding
Potential for breeding for resistance to physiological deterioration

The rapid post-harvest deterioration of cassava has been dealt with in a variety of ways by producers, processors and consumers (see Chapter Five). Because traditional management systems are well tailored to the crop's characteristics there has been little incentive historically to include reduced susceptibility to deterioration as a breeding objective. In addition, breeders have been reluctant to approach the problem because of the large influence of environmental growth conditions and associated preharvest stress (Table 7), limited genetic variability and a persistent but moderate negative correlation between low deterioration rate and high dry matter content (Kawano and Rojanaridpiched, 1983). However, with rapid urbanization occurring in all regions of the developing world and the increased potential of processed cassava products (see Introduction and Chapter Five), overcoming the post-harvest deterioration problems of cassava is becoming an important factor and has renewed interest in breeding for increased storage potential.

Cassava is vegetatively propagated by lignified stem cuttings and this has important implications for its evolution and genetic improvement. Clonally propagated crops have limited genetic plasticity and variations are slow to arise under natural conditions. However, cassava has an extensive adaptive plasticity and is able to tolerate a long dry season or irregular rainfall, low fertility, acid soils and various biological stresses. Cassava has no distinct period of physiological maturity. Harvest may be from six months to two years or more with an average of about one year. As cassava roots are not organs of propagation and can be harvested as needed, selection for postharvest conservation has not been a requirement. Traditional crop cultivars were selected by farmers to tolerate environmental and biological variations and to make efficient use of limited inputs. Cassava's rusticity, harvest flexibility and high productivity have combined to make it an enduring component of a broad range of tropical and subtropical cropping systems (Cock, 1985).

TABLE 7 Evaluation of susceptibility to physiological deterioration of 26 cultivars and hybrids harvested in Colombia (% deterioration)

Source: Wheatley, 1982.

Cassava has evolved primarily in adaptation to low input subsistence farming systems where yield per unit area was not a major concern, but, although still within essentially similar farming systems, cassava is now moving towards greater commercialization (see Chapter Five). The changes that call for new genotypes are the expectation of higher unit-area yields, increases in the pest and disease problems that result from more intensive cultivation (higher plant density, shorter rotations, monocropping) and the need for new quality traits for diversified markets (Henry, 1991). In most cassava producing areas the local cultivars are well suited to traditional markets and have the characteristics required for end product usage. These quality traits have been among those to receive major emphasis from farmer-breeders; in many cases maintaining existing quality characteristics as other traits are modified is one of the most challenging objectives to plant breeders.

Assembling and characterizing existing genetic diversity

The greatest genetic diversity for cassava exists in Latin America, although substantial diversification has taken place in Africa since the crop was introduced. Asia has the least diversity, with some countries such as Thailand and China producing cassava over large areas from relatively few cultivars. Most countries where cassava is a major crop have established improvement research programmes. In 1968, breeding programmes received a significant impetus with the establishment of two centres within the Consultative Group on International Agricultural Research (CGIAR) system -CIAT, founded in Cali, Colombia and the International Institute of Tropical Agriculture (IITA), established in Ibadan, Nigeria. The development of an international collection at CIAT in 1969-70 provided a wide range of genetic diversity from a single source (2 800 accessions) that could serve to augment the material available to breeding programmes. Safe and reliable long-term conservation of cassava germplasm has been of prime importance and CIAT's collection is maintained both in vitro and in the field.

In the last decade, with the assistance of the International Bureau of Plant Genetic Resources (IBPGR), CIAT's germplasm collection has expanded to over 5 000 accessions. These have been characterized for morphological traits according to the IBPGR descriptors (Gulick, Hershey and Esquinas-Alcazar, 1983), which are used as a means of genotype identification. The germplasm collection has also been agronomically characterized for yield, quality, pest and disease resistance and adaptation to a range of edaphoclimatic conditions. Other specific quality traits are often complex, costly and time-consuming procedures if entire germplasm collections are to be evaluated. However, evaluation of a core group which is a fraction of the total collection (5 to 10 percent) and is representive of the total genetic diversity (Brown, 1989) can provide overall indicators of genetic diversity at a fraction of the cost. A cassava core collection was recently established at CIAT (630 accessions) and is being evaluated for a range of traits, including cyanic acid and starch content, photosynthetic rate and nutrient use efficiency (Hershey et al., 1993). Banding patterns for alpha beta esterase isozymes and DNA fingerprinting have also been used for varietal fingerprinting to identify duplicates and as a means of assessing and describing genetic diversity (CIAT, 1991; Ocampo et al., 1993). A highly saturated molecular and physical map of cassava is being developed using random genomic and complementary DNA libraries (Angel et al., 1993).

Wild species of Manihot in their native habitat are restricted to Latin America where 98 species have been described (Rogers and Appan, 1973). Within the genus, only cassava is cultivated as a food crop and wild species have received little research attention. Field collections of wild species are being established and characterized at CIAT and IITA (with 29 and 33 species respectively). Isozyme patterns are being analysed to fingerprint the wild accessions and to estimate the relationships among species and between wild and cultivated germplasm (CIAT, 1993). The hypothesis that there is extensive natural introgression of wild species genes into cultivated cassava (Nasser, 1989) has yet to be confirmed, but a few cultivated genotypes show distinct wild type characteristics. Taxonomic aspects of the wild species have been studied by the Centro Nacional de Recursos Geneticos (CENARGEN) in Brasilia and crossing studies have been undertaken at IITA. A wild species native to northern Mexico and southwestern United States, M. walkerai, is reported to have roots with adventitious buds that can be used for propagation. The physiology of this species needs to be investigated and this characteristic confirmed.

Evolutionary relationships among wild species of Manihot (section Parvibracteatae) from Mexico and Central America have been studied using chloroplast and ribosomal DNA analysis, complemented by biogeographical and morphological data (Bertram, 1993). From this study it appeared that several wild species, including M. carthaginesis (section Carthaginensis) and M. aesculifolia, were genetically close to cassava. In addition to being morphologically similar to cassava, DNA results show evidence that M. aesculifolia is the closest extant relative of the crop.

Although additional input is required to evaluate wild species from South America, Bertram's studies ( 1993) already provide an informative basis for evaluating the use of wild species in cassava improvement breeding programmes. The characterization of the genetic variability in wild species will provide valuable information on cassava's evolution and help to identify new sources of specific traits. Due to their potential importance in breeding programmes expanded eftorts to collect, preserve end characterize the Manihot wild species germplasm are needed.

Genetics and cytogenetics

Like all wild species of Manihot studied to date, cassava has 36 chromosomes (Bai, 1987). Regular bivalent formation has been reported in the pollen mother cells, with few meiotic abnormalities. The completely paired pachytene bivalents vary in length from 19.3 to 40.0 microns. The haploid chromosomal complement has three functional nucleolar chromosomes and six chromosomal types represented in duplicate. This information has been used to suggest that the present-day cultivated types are allopolyploids of crosses between two closely related forms. Their two basic diploid parental taxa (x=9)7 while possessing six chromosomal types in common, differ in three chromosomes of their complement. Hence the present-day cultivars may be considered as segmental allopolyploids (Magoon, Krishnan and Bai, 1969). Similar pachytene studies have been carried out on M. glaziovii and a comparison with the karyotype of cassava showed many common features, including the same number and a similar morphology of chromosomes (Krishnan, Magoon and Bai, 1970). Studies on the genetics of cassava have been very limited and breeders have concentrated on obtaining the basic information required for effective genetic improvement of the crop. Attention is only beginning to focus on a more complete genomic characterization of the species and its relatives.

An extensive diversity exists for most traits examined to date. This may be due to introgression of wild species germplasm and to the many environments and uses for which cassava has been selected over thousands of years. Precise genetic control has been characterized for relatively few traits in cassava. Singlegene control has been demonstrated for leaf lobe width root surface colour, albinism, stem collenchyma colour, stem growth habit, root flesh pigmentation and male sterility (Hershey and Ocampo, 1989). There are no confirmed examples of physiological specialization on a gene-for-gene basis for pests or pathogens. Abroad range of agronomically important traits have been studied for their inheritance patterns. Results to date indicate that nearly all these traits are under multigenic control, with a high proportion of additive genetic effects (Iglesias and Hershey, 1994). Genetic variability within cassava for some traits, such as resistance to postharvest deterioration may be limited for breeding objectives. The application of molecular techniques, such as gene tagging and the identification of gene products, will complement conventional approaches to genetic studies of agronomically important traits.

Vegetative propagation is genetically a rather conservative strategy. New variation, resulting from natural crosses in multiclonal fields and possibly some introgression from wild species, is slow to arise. Plants originating from seed in the first generation following a cross are likely to be weak and at a distinct competitive disadvantage. New diversity for breeding programmes was initially obtained by the introduction of existing landrace cultivars from other regions, and the development of in vitro techniques has expanded the exchange of disease-free germplasm. However, superior landrace cultivars are unlikely to combine all the requirements of diverse production and market needs. Consequently, most breeding programmes look to recombination as the principal means of cultivar development.

Creation of new genetic diversity

The first significant large-scale creation of new diversity occurred through controlled crossing carried out by plant breeders. Most of the breeding methods applicable to outcrossing species can be used for cassava. Hybridization in cassava is relatively easy (Kawano, 1980) and open pollination schemes are extensively used to increase the amount of hybrid seed produced. In some regions the principal constraint to crossing is shy flowering and to date practical methods have not been developed to deal with non-flowering types. Flowering is controlled by the complex interaction of a range of genetic and environmental factors. In some areas cassava will flower all year long while in other locations flowering is seasonal. Cassava is monoecious, with pistillate flowers opening about two weeks before staminate flowers. Normally, few seeds are obtained (an average of I to 1.5) and acquiring large numbers of seeds is a labourintensive and tedious process. Most genotypes appear to suffer drastic inbreeding depression. Vegetative propagation to preserve superior heterozygotes greatly simplifies breeding and, whichever breeding method is used, heterozygocity needs either to be maintained or restored prior to subsequent vegetative propagation. Most programmes use some form of recurrent selection appropriate for the accumulation of many genes of minor effect. All existing commercial cultivars of cassava are probably highly heterozygous.

Although difficulties occur at the time of flowering, IITA has performed extensive intercrossing among wild species and between cassava and wild species. M. glaziovii has been used as a source of African mosaic-virus resistance in early East African breeding programmes (Nichols, 1947).

Mutation has been tried sparingly for cassava. One constraint is the need for selfing to achieve expression of recessive mutations occurring in the heterozygous state. This is very difficult for a large number of genotypes due to the asynchronous opening of staminate and pistillate flowers. Mutation could have more practical applications when haploid cells, such as microspores, can be regenerated into plantlets. Haploids and doubled haploids will be a significant research tool for cassava, with possible applications in genetics, evolution studies, expression of recessive genes and in a breeding system for true cassava seed. Experiments have been conducted on in vitro pollen germination, pollen tube growth and isolation, culture of young zygotic embryos, microsporogenesis, tetradstage microspores and microcalli formation (Cataño et al., 1993).

Polyploid induction, through colchicine-induced tetraploids, has been a subject of considerable research in India (Graner, 1941; Magoon et al., 1969). The clones produced generally exhibited the gigas characters associated with polyploidy, such as increases in leaf breadth and thickness, stomata! size, length and girth of petiole and flower size. Pollen sterility was high, but fertile pollen grains were much larger in size ( 180 to 196 microns) compared to diploids ( 125 to 140 microns). Considerable genotypic variation was demonstrated in response to polyploidy. Some clones became weak and could not be maintained, while others were maintained easily for several generations. Improved yield potential from polyploidy has not been found to be promising in any of the programmes. In India it was reported that a 42 percent increase in protein content was achieved by polyploid induction which subsequently disappeared with continued vegetative propagation.

At IITA work focused on changes in ploidy through inducing production of unreduced gametes, mainly through interspecific crosses. Four spontaneous tetraploids and two triploids were isolated from crosses between M. pruinosa or M. glaziovii and cassava. A majority of the interspecific crosses produced diploid pollen, but their frequencies varied with cross-combinations and with genotype within respective crosscombinations (Hahn, Bai end Asiedu, 1990). The presence of multivalents in the polyploids suggests that pairing and crossing over are taking place between cassava and its related Manihot species. There are no strong incompatibility barriers to interspecific hybridization, though considerable selection of parent cassava clones is necessary, which could indicate a highly fluid gene pool within the genus Manihot and likely widespread introgression among species.

Virtually no somaclonal variation has been detected from extensive studies on plants propagated from meristem culture or from somatic embryos (Szabados, Hoyos and Roca, 1987). Since either of these processes may result in considerable variation in many species, this result suggests that cassava is genetically very stable at these levels of tissue organization. Regeneration of cassava from individual cells is not possible at present and it is not known whether genetic stability would be similar for regenerated single cells.

Objectives in cassava breeding

Cassava is a crop that prospers in difficult and variable environments and breeders are faced with the need to consider many characters, each with multigenic control (Kawano et al., 1978). Cassava has been subject to small, incremental improvements at the hands of farmer-breeders for centuries but has received limited attention from modern breeding approaches. Recently improvement of this crop has been primarily oriented towards relatively few traits, most prominently to yield and pest and disease resistance. The former is a complex trait with many individual components, each involving numerous biochemical pathways and progress has been most pronounced in those cases where objectives were relatively simple. Resistance is normally determined by fewer genes, allowing for qualitative progress in a breeding programme. Advances in yield potential have been substantial but mainly in more favourable environments. These gains are often less prominent when translated to farmers' conditions, but the products of steady improvement are beginning to have substantial impact (Hershey and Jennings, 1992). Improvements in cassava in terms of adaptation, resistance, productivity and other traits have not been exhausted.

Breeding objectives for cassava have distinct regional differences that are determined by locally encountered constraints. Some of the major diseases and pests of cassava, such as bacterial blight, mealybug and green mites, were introduced into Africa from Latin America. The African mosaic virus, historically the most devastating biological constraint of cassava in Africa, is of unknown origin (Hahn, Terry and Leuschner, 1980). The importance of cassava for food security in humid tropical Africa has justified the establishment of many national programmes where crop improvement has had high priority (Hahn et al., 1979). Breeding efforts in Africa have focused on incorporating disease and pest resistance while arthropod pests have been approached mainly through biological control. In recent years, there has been sufficient progress in disease and pest resistance for other breeding objectives, such as cyanide and dry matter content and plant architecture, to be pursued. Eating quality continues to be a critical breeding objective in Africa. Efforts are also being directed towards expanding cassava production into the highlands and semi-arid regions of Africa, which require breeding and selection for low temperature and drought tolerance respectively.

In recent years, breeding research has expanded significantly in Asia because of the strong economic growth associated with cassava (Bottema and Henry, 1990). As fewer biological constraints exist in Asia and a large proportion of production is destined for industrial uses, breeders are concentrating on improved unit area dry matter yield and starch content. In Latin America as cassava is grown in diverse soil and climatic conditions it is subject to a wide range of pests and diseases. Breeding goals tend to be much more regionally focused but, in general, include a broader range of factors than in either Africa or Asia. Decentralized breeding programmes are required for cassava in order to address the wide range of physical environments' regionally distinct biological constraints and diverse market requirements.

The provision of appropriate environments for efficient germplasm selection at CIAT uses a subdivision of the cassava growing world into distinct agroecological zones based primarily on temperature and rainfall patterns. Pest and disease patterns are largely dependent on these environmental factors and gene pools have been developed for adaptation to each zone (Hershey, 1984; CIAT, 1992). Breeding for resistance to constraints is fundamental to the development of each gene pool. The basis for genetic improvement is the identification of representative environments where the principal traits of interest are consistently expressed at levels appropriate for selection (Iglesias et al., 1994). The need for extensive field testing is likely to continue for all crops.

The tools at the disposal of plant breeders have changed dramatically along with general developments in the biological sciences. The beginnings of a theory of inheritance was the first step towards a dramatic increase in knowledge about the nature and the manipulation of inherited variation. This in turn led to the development and application of tools for more efficient breeding; assembling and characterizing genetic variation, creating new variation and selection. No set of biological tools has created greater expectations than those generated by achievements in the area of molecular, cellular and other technologies broadly described as biotechnology. The rapidly continuing development of these tools has brought about fundamental changes in crop improvement research and broadened the range of potential traits for genetic modification. In addition to traditional breeders' objectives there are challenges that are not strictly breeding in nature and that go beyond the scope of the established approaches. These, if achieved, would have a major impact on the utilization of cassava. Objectives such as improvement in the post-harvest storage potential of roots, the development of acyanogenic cassava and the modification of starch characteristics for specialized markets are such challenges.

Potential for breeding for resistance to physiological deterioration

Breeders at CIAT are re-evaluating a broad range of germplasm using a more refined technique than was previously used. In the locations evaluated to date some cassava clones had virtually no deterioration eight days after harvest. Genetic variability accounted for 52 percent of the total observed variability, indicating the possibility of progress in a selection procedure. In order to study the intrinsic biochemical processes responsible for physiological deterioration, CIAT initially proposes to build up genetic stocks by crossing clones that demonstrate the extremes of each deterioration step. The biochemical processes involved in the rapid deterioration of cassava are essentially woundhealing responses (see Chapter Two). These responses are well-known in many plant species and involve the activity of many genes. Traditional breeding approaches will therefore need to employ methods for manipulation of quantitatively inherited traits.

It is conceived possible that conventional breeding could make significant contributions to reducing post-harvest deterioration of cassava. This is, however, likely to occur gradually in small progressive improvements.

Contents - Previous - Next