A. G. O. Dixon

Cassava breeding: opportunities and challenges

Although cassava is a major food crop, its scientific breeding began only recently compared with other crops. Significant progress has been achieved, particularly in Asia where cassava is used mainly for industrial processes and no major biotic constraints affect its productivity. Cassava breeding faces several limitations that need to be addressed. The heterozygous nature of the crop and parental lines used to generate new segregating progenies makes it difficult to identify parents with good breeding values. Breeding so far has been mainly based on a mass phenotypic recurrent selection. There is very little knowledge on the inheritance of traits of agronomic relevance. Several approaches have been taken to overcome the constraints in the current methodologies for the genetic improvement of cassava. Evaluations at early stages of selection allow for estimates of general combining ability effect or breeding values of parental lines. Inbreeding by sequential self-pollination facilitates the identification of useful recessive traits, either already present in the Manihot gene pool or induced by mutagenesis.

Simple sequence repeat marker diversity in cassava landraces: genetic diversity and differentiation in an asexually propagated crop

Abstract. Cassava (Manihot esculenta) is an allogamous, vegetatively propagated, Neotropical crop that is also widely grown in tropical Africa and Southeast Asia. To elucidate genetic diversity and differentiation in the crops primary and secondary centers of diversity, and the forces shaping them, SSR marker variation was assessed at 67 loci in 283 accessions of cassava landraces from Africa (Tanzania and Nigeria) and the Neotropics (Brazil, Colombia, Peru, Venezuela, Guatemala, Mexico and Argentina). Average gene diversity (i.e., genetic diversity) was high in all countries, with an average heterozygosity of 0.5358 ± 0.1184. Although the highest was found in Brazilian and Colombian accessions, genetic diversity in Neotropical and African materials is comparable. Despite the low level of differentiation [Fst(theta) = 0.091 ± 0.005] found among country samples, sufficient genetic distance (1-proportion of shared alleles) existed between individual genotypes to separate African from Neotropical accessions and to reveal a more pronounced substructure in the African landraces. Forces shaping differences in allele frequency at SSR loci and possibly counterbalancing successive founder effects involve probably spontaneous recombination, as assessed by parent-offspring relationships, and farmer-selection for adaptation.

Gene-based microsatellites for cassava (Manihot esculenta Crantz): prevalence, polymorphisms, and cross-taxa utility

BackgroundCassava (Manihot esculenta Crantz), a starchy root crop grown in tropical and subtropical climates, is the sixth most important crop in the world after wheat, rice, maize, potato and barley. The repertoire of simple sequence repeat (SSR) markers for cassava is limited and warrants a need for a larger number of polymorphic SSRs for germplasm characterization and breeding applications.ResultsA total of 846 putative microsatellites were identified in silico from an 8,577 cassava unigene set with an average density of one SSR every 7 kb. One hundred and ninety-two candidate SSRs were screened for polymorphism among a panel of cassava cultivars from Africa, Latin America and Asia, four wild Manihot species as well as two other important taxa in the Euphorbiaceae, leafy spurge (Euphorbia esula) and castor bean (Ricinus communis). Of 168 markers with clean amplification products, 124 (73.8%) displayed polymorphism based on high resolution agarose gels. Of 85 EST-SSR markers screened, 80 (94.1%) amplified alleles from one or more wild species (M epruinosa, M glaziovii, M brachyandra, M tripartita) whereas 13 (15.3%) amplified alleles from castor bean and 9 (10.6%) amplified alleles from leafy spurge; hence nearly all markers were transferable to wild relatives of M esculenta while only a fraction was transferable to the more distantly related taxa. In a subset of 20 EST-SSRs assessed by fluorescence-based genotyping the number of alleles per locus ranged from 2 to 10 with an average of 4.55 per locus. These markers had a polymorphism information content (PIC) from 0.19 to 0.75 with an average value of 0.55 and showed genetic relationships consistent with existing information on these genotypes.ConclusionA set of 124 new, unique polymorphic EST-SSRs was developed and characterized which extends the repertoire of SSR markers for cultivated cassava and its wild relatives. The markers show high PIC values and therefore will be useful for cultivar identification, taxonomic studies, and genetic mapping. The study further shows that mining ESTs is a highly efficient strategy for polymorphism detection within the cultivated cassava gene pool.

Serial analysis of gene expression (SAGE) of host-plant resistance to the cassava mosaic disease (CMD).

Cassava mosaic disease (CMD) is a viral disease of the important tropical staple crop cassava (Manihot esculenta) and preferred management involves use of host–plant resistance. The best available resistance is controlled by a single dominant gene. Serial analysis of gene expression (SAGE) was used to analyze the gene expression pattern in a bulk of 40 each of CMD resistant and susceptible genotypes drawn from a gene mapping progeny. Messenger RNA used for the SAGE analysis came from plants that were exposed to heavy disease pressure over a period of 2 years in the field. A total of 12,786 tags were studied, divided into 5733 and 7053 tags from the resistant and susceptible genotypes, respectively. Tag annotation was by PCR amplification using the tag sequence as sense primer and 4000 cassava ESTs generated from the bulk of CMD resistant genotypes. Annotation of more than 30 differentially expressed tags revealed several genes expressed during systemic acquired resistance (SAR) in plants and other genes involved in cell-to-cell and cytoplasm-to-nucleus virus trafficking. Differential expression of the most abundantly expressed tag, corresponding to a beta-tubulin gene, was confirmed by Northern Analysis. RFLP analysis of the tags in the parents and bulks of the CMD mapping progeny revealed only one tag, a WRKY transcription factor, associated with the region bearing the dominant CMD gene.

Alternate hosts of African cassava mosaic virus and East African cassava mosaic Cameroon virus in Nigeria

Cassava mosaic disease (CMD) caused by African cassava mosaic virus (ACMV) and East African cassava mosaic Cameroon virus (EACMCV) is the major constraint to cassava production in Nigeria. Sequences of the DNA-A component of ACMV and EACMCV isolates from leguminous plant species (Senna occidentalis, Leucana leucocephala and Glycine max), castor oil plant (Ricinus communis), a weed host (Combretum confertum) and a wild species of cassava (Manihot glaziovii) were determined. All ACMV isolates from these hosts showed 96–98% nucleotide sequence identity with cassava isolates from West Africa. EACMCV was found only in four hosts (S. occidentalis, L. leucocephala, C. confertum, M. glaziovii), and sequences of these isolates showed 96–99% identity with cassava isolates from West Africa. These results provide definitive evidence for the natural occurrence of ACMV and EACMCV in plant species besides cassava.

SSR markers reveal genetic variation between improved cassava cultivars and landraces within a collection of Nigerian cassava germplasm

Thirty-one improved cultivars and five Nigerian landraces of cassava were assessed at genomic DNA level with 16 SSR primers for genetic diversity study. The minimum number of SSR primers that could readily be used for identification of the 36 cassava genotypes was also determined. For the genetic diversity study, the similarity coefficients generated between improved cultivars and Nigerian landraces ranged from 0.42 to 0.84, and 12 distinct DNA cluster groups were identified at 0.70 coefficients using Numerical Taxonomy and Multivariate Analysis System software package. For the genotype identification study, the 16 SSR primers were screened by their polymorphic information content (PIC) values. Five SSR primers that have PIC values between 0.50 and 0.67 were selected and further assessed using simple arithmetic progression combination method. The results obtained revealed a combination of these 5 primers from SSR primers collection at IITA that could readily distinguish the 36 cassava genotypes at 0.93 similarity coefficient. These five primers clustered the 36 cassavas into 16 groups at 0.70 similarity coefficient. Application of this few SSR primers would ultimately reduce the cost and time of research for genetic diversity and genotype identification studies for the genetic improvement program of cassava.

Status of Cassava Begomoviruses and Their New Natural Hosts in Nigeria

A diagnostic survey was conducted in 2002-03 to determine the status of cassava mosaic begomoviruses in Nigeria and to ascertain if the virulent Ugandan variant of East African cassava mosaic virus (EACMV-Ug2) was present. Of the 418 farms visited, 48% had cassava with moderately severe or severe symptoms, whereas 52% had cassava with mild symptoms. These distributions were at random. Of the 1,397 cassava leaf samples examined, 1,106 had symptoms. In polymerase chain reaction tests, 74.1% of the symptom-bearing samples tested positive for African cassava mosaic virus (ACMV) alone, 0.3% for EACMV alone, 24.4% for mixed infections by the two viruses, and 1.2% did not react with any of the primers used. The two viruses also were detected in 32% of the 291 symptomless plants and in the whitefly vector samples. EACMV-Ug2, Indian cassava mosaic virus, and South African cassava mosaic virus were not detected in any of the whitefly or leaf samples. Most farms had ACMV in single infection as well as in mixed infections with EACMV. Most doubly infected plants showed severe symptoms. Two biological variants of ACMV were identified based on symptom expression on cassava in the field. ACMV and EACMV were detected in the leguminous plant Senna occidentalis (L.) Link and the weed Combretum confertum Lams.; these are new natural hosts of the viruses. Although the virulent EACMV-Ug2 was not detected, the occurrence of variants of ACMV and a high proportion of mixed infections by ACMV and EACMV, which could result in recombination events such as the one that produced EACMV-Ug2, demands appropriate measures to safeguard cassava production in Nigeria.

Identification of levels of resistance to cassava root rot disease (Botryodiplodia theobromae) in African landraces and improved germplasm using in vitro inoculation method

Cassava root rot disease is an increasing problem in Africa where yield losses of about 80% have been recorded. We evaluated 290 African landraces and 306 improved genotypes from the germplasm collections of the International Institute of Tropical Agriculture (IITA), for sources of resistance using root slice laboratory assay. Disease severity was assessed quantitatively by direct percentage estimation (PS) and by use of a rating scale (RS). Both methods of assessment were compared for identification of variability in the germplasm, and genotypes were classified into response groups using an enlarged rank-sum method that combined the PS and RS assessments. The two scoring methods revealed continuous variation (P < 0.001) for resistance in the sets of germplasm. Disease assessments based on PS and RS were highly correlated in both the improved germplasm (r = 0.75) and the landraces (r = 0.72). Based on PS assessment, 50 improved genotypes (16.3%) and 53 landraces (18.3%) showed significantly lower disease scores than the resistant control. The rank-sum method separated each set of collections into highly resistant, resistant, moderately resistant, moderately susceptible, susceptible and highly susceptible groups. Fifty-nine improved genotypes (16.4%) and 61 African landraces (16.9%) were identified as either highly resistant or resistant. Generally, these genotypes exhibited resistance by limiting the growth of the pathogen (reduced amount of invaded surface area). This type of rate-reducing resistance is highly heritable and a quantitative trait which can be harnessed in breeding. Genotypes subsets were identified for further studies into the genetic basis of resistance to root rot disease.

Field evaluation of root rot disease and relationship between disease severity and yield in cassava

Reports of cassava root rot disease from different African countries have increased in recent times. Field studies were conducted from July 1998 to October 1999 to determine a reproducible disease assessment method that would allow the comparison of results from different locations and an evaluation of the relationship between disease severity and root yield. Single point disease assessments at 6, 9, 12 and 15 months after planting (MAP) were compared to multiple points assessment based on the area under a disease progress curve (AUDPC). Single point assessments at 12 and 15 MAP, and the AUDPC identified continuous variation ( p ≤0.01) among the genotypes. However, a consistent result across trials was obtained only with the assessment based on AUDPC. Root dry yield (DYLD) at 15 MAP showed a strong negative correlation with AUDPC ( r =−0.74). Regression analysis also confirmed the negative relationship between yield and root rot severity. The five genotypes compared were separated into resistant (91/02324, 30572 and 92/0427) and susceptible (92/0057 and TME-1) groups. It was concluded that root rot disease may cause significant yield loss; however, the magnitude of the yield loss will depend on the susceptibility of the cassava genotype.

Assessment of laboratory methods for evaluating cassava genotypes for resistance to root rot disease

Field evaluation of six cassava genotypes for resistance to root rot disease was compared with three rapid laboratory methods (whole root inoculation, root slice inoculation, and stem inoculation) for resistance screening. Both the field evaluation and the three laboratory methods separated the varieties into resistant and susceptible groups. Genotypes 30572 and 91/02324 were resistant while 92/0247, 92/0057 and TME-1 were susceptible. One genotype (30001) was not consistent in its reaction between field evaluation and laboratory assays. In the laboratory assays with three fungal pathogens, different pathogens varied in their levels of virulence on host genotypes. With the most virulent pathogen (Botryodiplodia theobromae), the majority of the genotypes reacted in the same way across trials with the root slice and whole root assays. Due to the good correlation between the whole root assay and the field results, we recommend this for the routine assessment of cassava resistance to root rot disease and for the analysis of virulence of pathogen isolates. However, because of the advantages in terms of economy of labour, space, time, quantity of root and inoculum required, the root slice assay could be used for the preliminary screening of large cassava accessions. The selected genotypes can then be further screened with the whole root inoculation method.