Postharvest physiological deterioration (PPD) is a serious abiotic stress in cassava that renders the roots unmarketable, thereby reducing the economic value of the crop. This study was undertaken to determine farmers? perception of PPD and identify cassava genotypes with delayed PPD that can be used for better shelf life improvement of cassava.

A Participatory Rural Appraisal was conducted among 490 respondents that comprised farmers, processors, marketers and consumers in 12 communities in Ghana using Focus Group Discussion and semi-structured interview. Subsequently, 150 cassava accessions were assembled and assessed for genetic variability for PPD and other economic traits. Genetic diversity was assessed using morphological and molecular markers. A total of 19 qualitative and 195 SNP markers were used for the diversity studies. The Mantel test was used to assess the relationship between morphological traits and SNP markers. The 150 cassava accessions were evaluated in an Alpha lattice design to identify genotypes with delayed PPD using Booth?s method. Forty genotypes exhibiting comparatively low deterioration, high yield and dry matter were selected and evaluated in three locations for two years to assess genotype by environment interaction effect (GxE) on the expression of PPD. Genetic parameters including genotypic and phenotypic coefficient of variation and broad sense heritability were estimated using variance components to determine the feasibility of improving PPD through selection. Five accessions were selected and used for physicochemical and functional analysis of cassava day zero and seven after harvest to assess physiological and chemical changes during storage. The PRA revealed postharvest losses as one of the major production constraints. Genetic diversity estimated showed a moderate to high diversity among the 150 genotypes. Ninety six per cent of the SNP markers were polymorphic with a mean PIC value of 0.297 suggesting that these markers could be used for diversity assessment in cassava. Using the Jaccard?s and Nei?s similarity matrices, a higher mean similarity coefficient was obtained for the morphological traits (0.54) than was observed for the SNP markers (0.42). Dendrograms generated from both morphological and molecular analyses clustered the genotypes into five and three clusters respectively although there were disagreements between them. The correlation coefficient based on the Mantel test though significant was low (r=0.38, p=0.001). Significant variability also existed among accessions for PPD suggesting there were genotypic differences among accessions which could be used for PPD improvement. Genotypes K25, SW/00/064 and DMA00/031 with PPD scores of 9.1%, 9.6% and 9.9% respectively were identified, however, no accession with complete tolerance or resistance to PPD was identified. Genotype by environment analysis indicated the significance of genotype, location and year effects on all the traits studied. Genotype was more important for dry matter content than the other main effects while location and year effects were more important for fresh root yield and postharvest physiological deterioration respectively. The GGE biplot identified genotypes for specific locations for all the traits. Genetic variability components estimated revealed higher PCV than their corresponding GCV for all the traits studied. Broad sense heritability estimate was low for PPD (>20%) indicating that this trait is not under genetic control.