Cassava cooking properties characterization using NIRS on fresh ground cassava – Updated dataset 2020-2023

Davrieux Fabrice, Londoño Luis Fernando, Moreno Jhon Larry, Ospina Maria Alejandra, Luna Jorge, Duarte Cristian, Xiaofei Zang, Fenstemaker Sean, Tran Thierry, Dufour Dominique. 2024. Cassava cooking properties characterization using NIRS on fresh ground cassava – Updated dataset 2020-2023. Saint-Pierre : CIRAD-RTB Breeding Project, 32 p.

Résumé : This report, written under the RTB Breeding project, explores the characterization of cassava cooking properties using Near-Infrared Spectroscopy (NIRS). The study, conducted over five years, is based on the analyses of 3049 cassava genotypes (5815 samples) originated from various breeding populations at the CIAT research centre and from experimental fields. The samples were evaluated for their water absorption capacity after 30 minutes of boiling (WA30, a key indicator of cooking quality) and their near infrared fingerprints. Genotypes were classified into 2 classes according to their WA30 values with C1 corresponding to WA30 < 12% and C2 corresponding to WA30 ≥ 12%. The repartition by classes was 3111 samples in C1 and 2704 samples in C2. To classify the genotypes, the study used Principal Components Analysis (PCA) to reduce the dimensionality of the spectral data, followed by Locally Weighted Partial Least Squares Regression Discriminant Analysis (LWPLSRDA) for predicting the genotype classes. The robustness and accuracy of these models were key focuses. The final LWPLSRDA model achieved a classification accuracy of 80% in distinguishing between "low cooking time" and "long cooking time" genotypes. The model was balanced in terms of sensitivity and specificity, which means a high capacity to detect true positives and true negatives genotypes. This study confirms that there is a relationship between the NIR fingerprint of fresh cassava tubers, largely due to chemical composition, and their cooking capacity. By integrating NIRS technology into cassava breeding, researchers can more efficiently identify varieties with optimal cooking properties, thereby supporting the development of superior cassava cultivars.

Mots-clés libres : Breeding, Cassava, Near Infrared Spectroscopy, LWPLSRDA, Classification, Cooking ability, Water absorption

Agences de financement hors UE : International Potato Center, Bill and Melinda Gates Foundation, Centre de Coopération Internationale en Recherche Agronomique pour le Développement

Projets sur financement : (FRA) RTB Breeding

Auteurs et affiliations

• Davrieux Fabrice, CIRAD-PERSYST-UMR Qualisud (REU)
• Londoño Luis Fernando, CIAT (COL)
• Moreno Jhon Larry, CIAT (COL)
• Ospina Maria Alejandra, CIAT (COL)
• Luna Jorge, CIAT (COL)
• Duarte Cristian, CIAT (COL)
• Xiaofei Zang, CIAT (COL)
• Fenstemaker Sean, CIAT (COL)
• Tran Thierry, CIRAD-PERSYST-UMR Qualisud (FRA) ORCID: 0000-0002-9557-3340
• Dufour Dominique, CIRAD-PERSYST-UMR Qualisud (FRA) ORCID: 0000-0002-7794-8671

Source : Cirad-Agritrop (https://agritrop.cirad.fr/610379/)

Métriques

[ Page générée et mise en cache le 2024-09-23 ]