Diversity and genetic distance are required as initial foundations to identify germplasm Indonesian cassava potential for food, industrial, and biofuel resources. This study used 181 cassava (Manihot esculenta Crantz.) accessions from all islands in Indonesia, i.e. Java, Sumatera, Kalimantan, Sulawesi, Maluku, Nusa Tenggara Timur and Papua Islands. The study was conducted in July 2013 to March 2014. Research experiment design was arranged in Augmanted Design with three control plants per row. There were traits of morphological rod and leaf as parameter, the number of 19 traits. The analysis was using Principal Component Analysis (PCA) and Agglomerative Hierarchical Clustering (AHC). Results of this study are genetic diversity and distance cassava from Indonesia with a wide diversity level of 49.82 % and from 1 to 17 genetic distance spread throughout Indonesia.

There is an abundance of cassava (Manihot esculenta Crantz.) genetic resources in Indonesia, and the local accessions are inseparable from the community of Indonesia. Several of the cultivars have cultural significance and over time have been bred for specific uses and products. The specific use and combination of traits encourages the use of local cultivars or aims for genetic improvement of the local cultivars. The objective of this study was to measure character variability and to categorize cassava clones based on specific characteristics to better inform selection criteria. A total of 156 cassava clones collected from all over Indonesia were evaluated along with three clones of the local cultivar Jatinangor as checks. This is basic research, so the data information can be a complement to the cassava germplasm in Indonesia. The experiment was performed as an augmented block design. The variability of characteristics was analyzed using principal components analysis with a Pearson correlation. Grouping of clones was accomplished using a symmetric biplot function. Three first principal components contributed to the maximum variability of cassava by 87.85 %, and characters that contributed variability had factor loadings>0.6. Having variability in characteristics suggests that there is an opportunity for performance-based clone selection. In this study,nine cassava clones with desirable trait combinations were identified based on PCA, of which four were identified as the best performing clones.