Native starch evaluation and genetic distance analysis using AFLP of elite cassava (Manihot esculenta Crantz) genotypes from Malawi

English:The starchy tuberous roots of cassava (Manihot esculenta Crantz) provide more than half of the calories consumed by more than 800 million people in the South America, Asia and Africa. Cassava is a staple for more than 30 % of the Malawi population, while in the rest of the country it is grown for food security, as a snack, and a cash crop. A prerequisite for any genetic improvement programme of cassava is knowledge of the extent of genetic variation present between cultivars and genetic distances between them, and between cassava and closely related species with which hybrids could be produced. This can be achieved through characterisation of germplasm either using morphological, biochemical or DNA markers. Industries that use starch in Malawi have not been willing to use cassava starch because the powder sold by some suppliers as 'cassava starch' was inferior. This study was therefore initiated with the following objectives: (1) To determine the genetic distances/relatedness of commercial Malawi cassava varieties and promising clones using AFLP and morphology descriptors; (2) To evaluate the Malawi commercial cassava varieties and promising clones for starch extraction and quality; (3) To assess the feasibility of using native cassava starch in the industrial sector in Malawi; and (4) To evaluate the commercial cassava varieties in Malawi and promising clones for dry matter content.Trials with 20 cassava genotypes were planted at Chitedze and Makoka in Malawi in November 2000, and were harvested in December 2001. Morphological characterisation was done during plant growth and at harvest. The quality parameters included: protein, moisture and ash content, pH and whiteness. Feasibility studies in the use of cassava starch in the pharmaceutical, textile, battery making and packaging industries started in December 2001. 106 The results show that all the cassava genotypes produced starch containing no detectable protein, similar to the starch used in the pharmaceutical industry. The moisture content ranged from 11.85 to 13.65 %, which is lower than the recommended maximum of 14 %. The recommended maximum ofO.5 % for ash was much higher than the values of cassava starch, which ranged from 0.10 to 0.20 %. The recommended pH for starch is between 4.5 and 7.0, and cassava starch was within these limits, ranging from 5.0 to 5.9. The cassava starch was as white as the corn starch currently being used. Trials have shown that cassava starch can successfully be used in the making of tablets, batteries, packaging material and textile manufacture. Thus, native cassava starch is suitable for use in various industries. This study has also dispelled fears that cassava starch is of low quality and that it is not effective in some industries. Hence, the ideal quality of starch was confirmed in this study. What is required is strict quality control for cassava starch so that industry receives what they pay for. MBS needs to develop national standards for starch, which will help in the enforcement of the quality control for any starch produced in Malawi. This task will be facilitated by the recommendations made in this thesis. Additive main effects and multiplicative interaction (AMMI) was strongly correlated with other stability parameters like Wi-ecovalence, stability variance - no covariate and cultivar superiority measure. AMMI is therefore recommended for use in the stability analysis of starch quality parameters since it provides additional information on the allocation of unstable genotypes into their appropriate environments. This study has demonstrated that genotype has a greater influence on root dry matter than the environment. This agrees with the hypothesis that one or a few major genes control root dry matter in cassava. This study has also revealed that the genetic distances for the local cultivars, locally bred clones and introduced genotypes are equally divergent. This is supported by the distribution of the genotypes in different clusters despite their origin Genetic distances determined by morphological characterisation correlates to similar values using AFLP fingerprinting (r=0.98). It is however, a prerequisite that morphological characterisation be based on constant traits, and the conversion of the morphologic data into binary characters needs careful consideration to meaningful results.