| dc.description.abstract | Cassava mosaic geminiviruses (CMG) and Cassava brown streak virus (CBSV) cause the highest yield losses in Africa's cassava production. Cassava brown streak disease (CBSD) in particular has been and continues to be a significant constraint to optimal cassava production in Eastern and Southern Africa. While CBSD has not been reported in West Africa, its recent rapid spread and damage to cassava productivity in Eastern, Central and Southern Africa is of great concern. This study was initiated to determine the status of the selected Nigerian cassava genotypes with respect to CBSD resistance. Such activity is required as a pre-breeding step that aims towards minimizing CBSD impact in Nigeria, in the event that the disease invades the country. Information generated from this step will inform on the level and extent of CBSD resistance breeding required for the Nigerian genotypes. Modern breeding techniques have taken center stage in cassava breeding, to help shorten breeding stages and cycles. One of such techniques is the use of genomic selection (GS) which promises to shorten cassava breeding cycles and increase genetic gain per year. An important aspect of GS is the prediction accuracy, which informs the breeder on whether the technique can be used to sufficiently improve genetic gain per cycle of a population, as against phenotypic selection. As such, the study herein aimed to contribute towards the development of improved Nigerian cassava varieties enhanced with CBSD resistance. Specifically, the study aimed to (i) determine the field reaction of Nigerian cassava populations to CBSD (ii) determine prediction accuracy of genomic selection for CBSD resistance in Nigerian cassava germplasm. To this effect, a total of 1980 full-sib cassava seedlings from 106 families were assessed in the field at seedling stage and a subset of 569 clones were selected and assessed at clonal stage in Namulonge, Central Uganda; a known hotspot for CBSD screening. Results indicated that foliar, root incidences and severities varied significantly (p≤0.01, p≤0.001) except for CBSD incidence at six months. Lowest and highest plot-based heritability estimates for CBSD were registered for CBSDrs (0.71) and CBSD6i (0.5). Positive and highly significant correlations were noted between CBSDri (root incidence) and CBSDrs (root severity) (r = 0.90***), CBSD6i (foliar incidence at six months) and CBSD6s (foliar severity at six months) (r = 0.89***), CBSD3i (foliar incidence at three months) and CBSD3s (foliar severity at three months) (r = 0.89***). Significant positive correlations were also noted between CBSD3s and CBSD6i (r = 0.77***), CBSD3s and CBSD6s (r = 0.78***), CBSD3s and CBSDri (r = 0.40***), and between CBSD3s and CBSDrs (r = 0.35***). On the other hand, Fresh root weight (FreshRW) negatively correlated with CBSD6i, CBSD6s, CBSDri and CBSDrs respectively (r = -0.52***, r = -0.45***, r = -0.21*** and r = -0.22***). Similarly, CBSD3s correlated negatively with CMD3i, CMD3s, CMD6i and CMD6s respectively (r = -0.25***, r = -0.25***, r = -0.20***, and r = -0.21***). Thirty CBSD tolerant cassava genotypes were selected using a non-weighted rank summation selection index. Conclusively, results revealed that the Nigerian accessions exhibited significant susceptibility to CBSD within the first two years of evaluation. This information will aid future breeding decisions for improvement of CBSD resistance among Nigerian cassava varieties. For the genomic prediction study, Cross validation within the Nigerian populations revealed that genomic best linear unbiased prediction (GBLUP) and reproducing kernel Hilbert spaces (RKHS) models (r = 0.20 - 0.22) performed better than BL and BRR (r = 0.16 - 0.18) in predicting CBSD response in the population. Prospects of between-population predictions were also tested whereby two Ugandan cassava populations (Cycle 0 (C0) and Cycle 1 (C1) genomic selection populations) from National Crops Resources Research Institute (NaCRRI), Namulonge were used to predict CBSD response in the Nigerian population. Briefly, the C0 populations consisted of 384 clones while C1 population had 638 clones. The Nigerian population consisted of 297 genotyped clones that remained after clonal evaluation. Using phenotypic and genotypic data (SNP markers) for the two Ugandan cassava populations as training sets, CBSD symptoms for the Nigerian clones were predicted. The C1 training set yielded relatively higher prediction accuracies (CBSD3s r = 0.13, CBSD6s r = 0.20, CBSDRs r = 0.02) than the C0 training set (CBSD3s r = -0.03, CBSD6s r = 0.04, CBSDRs r = -0.10) across all traits and prediction models. The generated results indicated that genomic selection could be utilized to improve breeding efficiency for CBSD resistance in the Nigerian population. Genomic selection would help fast track the breeding process for developing CBSD resistant/tolerant Nigerian cassava varieties. | en_US |
