Integrating end-user preferences into breeding programmes for roots, tubers and bananas.

"Consumers have their say: assessing preferred quality traits of roots, tubers and cooking bananas, and implications for breeding" special issue, brings together new knowledge about quality traits required for roots, tubers and bananas (RTB) varieties to successfully meet diverse user preferences and expectations, along the variety development and RTB value chains (production, processing, marketing, food preparation, consumption). Key RTB crops in sub-Saharan Africa are cassava, yams, sweetpotatoes, potatoes and bananas/plantains. They are mainly consumed directly as boiled pieces or pounded in the form of smooth, not sticky, and stretchable dough. They are also stewed, steamed or fried. Cassava, the most widely grown RTB, is generally boiled, stewed or steamed in Eastern and Southern Africa, and in West and Central Africa is usually processed directly into derivative products, e.g. whole root fermentation through retting or heap fermentation; fermentation/dewatering of the mash. Biophysical and social knowledge presented in this issue help elaborate goals for both the processing unit operations (food scientist control) and variety traits (breeder control).

Association genetics of bunch weight and its component traits in East African highland banana (Musa spp. AAA group).

KEY MESSAGE: The major quantitative trait loci associated with bunch weight and its component traits in the East African highland banana-breeding population are located on chromosome 3. Bunch weight increase is one of the major objectives of banana improvement programs, but little is known about the loci controlling bunch weight and its component traits. Here we report for the first time some genomic loci associated with bunch weight and its component traits in banana as revealed through a genome-wide association study. A banana-breeding population of 307 genotypes varying in ploidy was phenotyped in three locations under different environmental conditions, and data were collected on bunch weight, number of hands and fruits; fruit length and circumference; and diameter of both fruit and pulp for three crop cycles. The population was genotyped with genotyping by sequencing and 27,178 single nucleotide polymorphisms (SNPs) were generated. The association between SNPs and the best linear unbiased predictors of traits was performed with TASSEL v5 using a mixed linear model accounting for population structure and kinship. Using Bonferroni correction, false discovery rate, and long-range linkage disequilibrium (LD), 25 genomic loci were identified with significant SNPs and most were localized on chromosome 3. Most SNPs were located in genes encoding uncharacterized and hypothetical proteins, but some mapped to transcription factors and genes involved in cell cycle regulation. Inter-chromosomal LD of SNPs was present in the population, but none of the SNPs were significantly associated with the traits. The clustering of significant SNPs on chromosome 3 supported our hypothesis that fruit filling in this population was under control of a few quantitative trait loci with major effects.

Genomic Prediction in a Multiploid Crop: Genotype by Environment Interaction and Allele Dosage Effects on Predictive Ability in Banana.

Improving the efficiency of selection in conventional crossbreeding is a major priority in banana ( spp.) breeding. Routine application of classical marker assisted selection (MAS) is lagging in banana due to limitations in MAS tools. Genomic selection (GS) based on genomic prediction models can address some limitations of classical MAS, but the use of GS in banana has not been reported to date. The aim of this study was to evaluate the predictive ability of six genomic prediction models for 15 traits in a multi-ploidy training population. The population consisted of 307 banana genotypes phenotyped under low and high input field management conditions for two crop cycles. The single nucleotide polymorphism (SNP) markers used to fit the models were obtained from genotyping by sequencing (GBS) data. Models that account for additive genetic effects provided better predictions with 12 out of 15 traits. The performance of BayesB model was superior to other models particularly on fruit filling and fruit bunch traits. Models that included averaged environment data were more robust in trait prediction even with a reduced number of markers. Accounting for allele dosage in SNP markers (AD-SNP) reduced predictive ability relative to traditional bi-allelic SNP (BA-SNP), but the prediction trend remained the same across traits. The high predictive values (0.47- 0.75) of fruit filling and fruit bunch traits show the potential of genomic prediction to increase selection efficiency in banana breeding.

Trait variation and genetic diversity in a banana genomic selection training population.

Banana (Musa spp.) is an important crop in the African Great Lakes region in terms of income and food security, with the highest per capita consumption worldwide. Pests, diseases and climate change hamper sustainable production of bananas. New breeding tools with increased crossbreeding efficiency are being investigated to breed for resistant, high yielding hybrids of East African Highland banana (EAHB). These include genomic selection (GS), which will benefit breeding through increased genetic gain per unit time. Understanding trait variation and the correlation among economically important traits is an essential first step in the development and selection of suitable GS models for banana. In this study, we tested the hypothesis that trait variations in bananas are not affected by cross combination, cycle, field management and their interaction with genotype. A training population created using EAHB breeding material and its progeny was phenotyped in two contrasting conditions. A high level of correlation among vegetative and yield related traits was observed. Therefore, genomic selection models could be developed for traits that are easily measured. It is likely that the predictive ability of traits that are difficult to phenotype will be similar to less difficult traits they are highly correlated with. Genotype response to cycle and field management practices varied greatly with respect to traits. Yield related traits accounted for 31-35% of principal component variation under low and high input field management conditions. Resistance to Black Sigatoka was stable across cycles but varied under different field management depending on the genotype. The best cross combination was 1201K-1xSH3217 based on selection response (R) of hybrids. Genotyping using simple sequence repeat (SSR) markers revealed that the training population was genetically diverse, reflecting a complex pedigree background, which was mostly influenced by the male parents.

The triploid East African Highland Banana (EAHB) genepool is genetically uniform arising from a single ancestral clone that underwent population expansion by vegetative propagation.

KEY MESSAGE: All East African Highland Banana varieties are genetically uniform having arisen from a single clone introduced to Africa. East African Highland bananas (EAHBs) are a subgroup of triploid (AAA genome) bananas of importance to food security in the Great Lakes region of Africa. Little is known about their genetic variation, population structure and evolutionary history. Ninety phenotypically diverse EAHB cultivars were genotyped at 100 SSR microsatellite markers to investigate population genetic diversity, the correlation of genetic variability with morphological classes, and evolutionary origins since introduction to Africa. Population-level statistics were compared to those for plantain (AAB) and dessert (AAA) cultivars representing other M. acuminata subgroups. EAHBs displayed minimal genetic variation and are largely genetically uniform, irrespective of whether they were derived from the distinct Ugandan or Kenyan germplasm collections. No association was observed between EAHB genetic diversity and currently employed morphological taxonomic systems for EAHB germplasm. Population size dynamics indicated that triploid EAHBs arose as a single hybridization event, which generated a genetic bottleneck during foundation of the EAHB genepool. As EAHB triploids are sterile, subsequent asexual vegetative propagation of EAHBs allowed a recent rapid expansion in population size. This provided a basis for emergence of genetically near-isogenic somatic mutants selected across farmers and environments in East Africa over the past 2000 years since EAHBs were first introduced to the African continent.

Transgenic expression of the rice Xa21 pattern-recognition receptor in banana (Musa sp.) confers resistance to Xanthomonas campestris pv. musacearum.

Banana Xanthomonas wilt (BXW), caused by the bacterium Xanthomonas campestris pv. musacearum (Xcm), is the most devastating disease of banana in east and central Africa. The spread of BXW threatens the livelihood of millions of African farmers who depend on banana for food security and income. There are no commercial chemicals, biocontrol agents or resistant cultivars available to control BXW. Here, we take advantage of the robust resistance conferred by the rice pattern-recognition receptor (PRR), XA21, to the rice pathogen Xanthomonas oryzae pv. oryzae (Xoo). We identified a set of genes required for activation of Xa21-mediated immunity (rax) that were conserved in both Xoo and Xcm. Based on the conservation, we hypothesized that intergeneric transfer of Xa21 would confer resistance to Xcm. We evaluated 25 transgenic lines of the banana cultivar 'Gonja manjaya' (AAB) using a rapid bioassay and 12 transgenic lines in the glasshouse for resistance against Xcm. About 50% of the transgenic lines showed complete resistance to Xcm in both assays. In contrast, all of the nontransgenic control plants showed severe symptoms that progressed to complete wilting. These results indicate that the constitutive expression of the rice Xa21 gene in banana results in enhanced resistance against Xcm. Furthermore, this work demonstrates the feasibility of PRR gene transfer between monocotyledonous species and provides a valuable new tool for controlling the BXW pandemic of banana, a staple food for 100 million people in east Africa.

Sequence characteristics of potato virus Y recombinants.

Potato virus Y (PVY) is one of the most economically important plant pathogens. The PVY genome has a high degree of genetic variability and is also subject to recombination. New recombinants have been reported in many countries since the 1980s, but the origin of these recombinant strains and the physical and evolutionary mechanisms driving their emergence are not clear at the moment. The replicase-mediated template-switching model is considered the most likely mechanism for forming new RNA virus recombinants. Two factors, RNA secondary structure (especially stem-loop structures) and AU-rich regions, have been reported to affect recombination in this model. In this study, we investigated the influence of these two factors on PVY recombination from two perspectives: their distribution along the whole genome and differences between regions flanking the recombination junctions (RJs). Based on their distributions, only a few identified RJs in PVY genomes were located in lower negative FORS-D, i.e. having greater secondary-structure potential and higher AU-content regions, but most RJs had more negative FORS-D values upstream and/or higher AU content downstream. Our whole-genome analyses showed that RNA secondary structures and/or AU-rich regions at some sites may have affected PVY recombination, but in general they were not the main forces driving PVY recombination.

NE-11 represents a new strain variant class of Potato virus Y.

This report describes the characterization by whole-genome sequencing of four PVY isolates with unique combinations of molecular and symptomatic characteristics. Three of these four isolates were of type PVY(N:O) (ID-1, OR-1, PN10A), including one of "type B", which contains an extra recombination event in the 5'UTR/P1 cistron; the other (NE-11) represents a novel PVY molecular genotype, previously misclassified as a PVY(NA-NTN) isolate. The full genome sequence of this latter isolate is unique inasmuch as it is nearly identical to that of PVY(N) isolates for the first 2,000 nucleotides (nts), after which it very strongly resembles PVY(NA-NTN) isolates for the next 600 nts. For the final 7,000 nts of its genome, NE-11 shares intermediate identity with these other two previously reported classes of PVY(N) genomes, except for a portion of the capsid protein region in which it resembles neither. Recombination in each of the four isolates was verified by a suite of recombination detection programs. PN10A represents the first complete sequence of a PVY strain variant of the class reported as PVY(N)-W (or PVY(N:O)) type B. Specific PCR assays for two unique regions of NE-11 are presented that will allow the identification of this strain variant by other researchers.