Abstract

Cassava (Manihot esculenta Crantz) is a key staple in the Democratic Republic of  Congo (DRC), yet the diversity of farmer-maintained cultivars remains insufficiently documented, particularly in the southwest. This study assessed the morphological diversity and population structure of cassava cultivars collected from farmers’ fields in this region. A survey was conducted in 112 fields across Kinshasa, Kongo Central, and Mai-Ndombe, where 220 cultivars were recorded. Characterization relied on nine qualitative descriptors from standard cassava lists, complemented by altitude and whitefly abundance. Multiple Correspondence Analysis (MCA) was used to explore phenotypic variability, followed by Hierarchical Ascending Classification (HAC) to group cultivars based on similarity. Discriminant Factor Analysis (DFA) evaluated the robustness of the classification. The first five MCA dimensions explained 42.9% of total variation. Leaf pigmentation, petiole colour, plant age, and cassava mosaic disease severity were the main discriminating traits. HAC identified seven phenotypic groups of unequal size, indicating substantial diversity within the population. DFA showed high classification accuracy (94.1%), confirmed by Leave-One-Out cross-validation (92.3%). Overall, the results reveal significant phenotypic variability in cassava cultivated in southwestern DRC and emphasize its importance for breeding programs aimed at enhancing disease resistance, productivity, and adaptation to local environmental conditions. These findings provide a valuable baseline for conservation and genetic improvement efforts and highlight the need for integrating farmer knowledge with formal breeding strategies. Future research should incorporate molecular markers to better resolve genetic relationships and support the development of resilient, high-yielding cultivars adapted to diverse agroecological conditions in practice.

This work is licensed under a Creative Commons Attribution 4.0 License.