| dc.description.abstract | Virus diseases are a major constraint to sweetpotato production in East Africa. The most important is the Sweet potato virus disease (SPVD), a result of co-infection of Sweet potato chlorotic stunt virus (SPCSV) and Sweet potato feathery mottle virus (SPFMV). Studies were done on different aspects of resistance to SPVD, and to determine the presence of Sweet potato virus G (SPVG), Sweet potato virus 2 (SPV2), and Sweet potato leaf curl virus (SPLCV), viruses that have not been reported infecting sweetpotato in Kenya. None of the samples reacted to antisera for either SPVG or SPV-2. SPLCV was detected infecting sweetpotato in Kenya for the first time. Some sweetpotato genotypes have irregular distribution or low virus titers, or recover from SPVD. The possibility of using this resistance to obtain virus free cuttings from field-grown sweetpotato vines for propagation was studied. Vines were cut into three pieces (15 cm, 15-30 cm, and >30 cm from the apex) and tested for SPCSV, SPFMV and Sweet potato mild mottle virus (SPMMV), the most common viruses in Kenyan fields. The viruses were equally present in all sections of infected vines and no section was any more likely to be virus free. Accumulation patterns of SPCSV and SPFMV in mixed infections were compared in SPVD-susceptible cultivars, ‘Beauregard’ and ‘Namaswakhe’, and resistant cultivars, ‘Naspot I’ and ‘Mar Ooko’. Virus titers were estimated using real-time quantitative PCR. Resistance in ‘Naspot I’ and ‘Mar Ooko’ was associated with reduced SPCSV and SPFMV multiplication, respectively. Titers of both viruses increase to certain thresholds after which symptoms appear, indicating that both viruses are important in SPVD development. To determine if SPVD resistant genotypes could be identified using molecular markers, sweetpotato genotypes were selected and classified as resistant or susceptible and amplified fragment length polymorphism (AFLP) marker profiles used in association studies. Analysis of molecular variance found significant (P<0.002) differences between the two groups. Discriminant and logistic regression analysis were used to select informative markers, and to develop models to classify the two groups. Four markers, which gave 94% correct classification of a test population, were selected by both statistical methods. | en |
