**4. Conclusion**

*Legume Crops – Characterization and Breeding for Improved Food Security*

top of the agenda.

II.Nutrition-sensitive trait selection for improved dietary quality: interest in cowpea's nutrition quality in Africa is an old issue [76], but it has been overlooked over the years. However, with the continent's nutrition agenda becoming increasingly important, a targeted breeding agenda on the nutritional quality of the crop is needed [12, 77]. Contrarily to Africa, the production of varieties with high dietary quality has deserved much attention in Europe and Asia [78, 79]. Currently, screening segregating populations for traits such as Fe, Zn, Cu and Mo content is in progress [80]. Such efforts are essential to improve the crop's contribution to local diets, as well as for the establishment of nutrition sensitive food systems. Special attention should be also given to fresh leaves and pods rather than solely focusing on dry pods and grain as it has happened so far. Increasing protein and mineral content, the latter also through biofortification, needs to be on

III.Breeding approach: most breeding programs in Africa rely on open environment conventional breeding technics centered mostly on single trait selection methods. However, in the developing world, molecular characterization of germplasm, based in modern genomics and molecular marker-assisted selection [1] and genetic engineering [80, 81] coupled with digital imaging in high-throughput phenotyping [82], historical data [83, 84] and model-assisted selection [84–87] have revolutionized crop breeding programs. Such approach facilitated molecular, morpho-agronomic, physiological and biochemical characterization of cowpea germplasm to identify the best performing genotypes [88]. This integrative screening and selection approach represents a clear shift from single-trait to multipletrait selection [85], something that is scantly done in African screening programs. By doing multiple-trait selection, the effectiveness and efficiency of breeding programs have been significantly improved in Europe, America and Australia, where significant investments in research infrastructure and human resource training has been made [89]. Model assisted breeding has proved to be fundamental in helping underpin prediction of likely phenotypical consequences of trait and genetic variations in targeted environments [86]. Furthermore, the agricultural production simulator (APSIM) has been successfully used in phenotyping and evaluating Genotype × Environment × Management (G × E × M) effects on drought adaptation. The growing interest in genotype-to-phenotype (G2P) models which predict phenotypic traits as a function of genotypic and environmental inputs is currently helping to enhance phenotype screening [89]. Additionally, the use of speed breeding chambers (SBC) [90], is also a recent and important advance in breeding programs. Such facilities allowed breeders to achieve up to six generations per year from spring wheat, durum wheat, barley, pea, chickpea and groundnuts, instead of one to three generations per year usually possible under field conditions and glasshouse, respectively [91].

IV.Improve cross-country coordination mechanisms and systematization of existing information: over the last decades several projects involving cowpea landraces screening and the assessment of their genetic diversity have been conducted in Africa [77]. However, the knowledge generated from this research is scattered all over the region and needs to be systematized and made available to aid current and future breading programs. For that to happen, cross-country coordination mechanisms and collaborative research

opportunities need to be improved.

**102**

With an increasing world population, there is an urgent need to re-structure the R&D agenda in SSA towards the development of elite crop varieties that are more likely to successfully cope with future climate conditions. Cowpea, despite its high plasticity to survive in harsh environments, will not be an exception. The crop's importance in SSA as a food crop, animal feed and nutrient cycling agent makes it a candidate crop for future improvement and to operationalize the continents' nutrition agenda. For that, coordinate R&D efforts should be made at the regional level, in order to: (i) address the best production and breeding practices, through a wide screening of landraces towards the identification of the best performing genotypes (yield and nutritional quality) under limiting environmental conditions; (ii) identify multiple breeding traits and molecular tools for marker-assisted selection; and (iii) develop fast and reliable methods for variety certification, linked to important investment in R&D facilities and advanced training of human resources.
