RESUMEN
Efforts are currently underway to improve carotenoids content in cassava roots through conventional breeding as a strategy to reduce vitamin A deficiency. However, only few samples can be quantified each day for total carotenoids (TCC) and ß-carotene (TBC) contents, limiting the gains from breeding. A database with >3000 samples was used to evaluate the potential of NIRS and chromameter devices to predict root quality traits. Maximum TTC and TBC were up to 25.5 and 16.6 µg/g (fresh weight basis), respectively. NIRS predictions were highly satisfactory for dry matter content (DMC, R(2): 0.96), TCC (R(2): 0.92) and TBC (R(2): 0.93). NIRS could also distinguish roots with high or low cyanogenic potential (R(2): 0.86). Hunter color parameters could also be used for predictions, but with lower accuracy than NIRS. NIRS or chromameter improve selection protocols, allowing faster gains from breeding. Results also demonstrate that TBC and DMC can be improved simultaneously (required for the adoption of biofortified cassava).