Skip to Main content Skip to Navigation
Conference papers

Étude du continuum pratiques d'élevage des bovins – carcasse – muscle – viande pour une gestion optimale des qualités des produits : stratégies statistiques applicables pour l’analyse de métadonnées

Abstract : Animal production experiments generate large numbers of individual data called metadata. However, these have not been extensively used in animal sciences including meat science compared to other disciplines. Therefore, our project intend to use meat science metadata that spread over the continuum, from farmgate-to-meat, to identify how can we jointly manage carcass and beef qualities using rearing practices applied during the whole life of the animals. To achieve this challenging objective, we implemented several statistical strategies to analyze the metadata of the continuum from farmgate-to-meat. In this paper, we will present two examples of our current results obtained using two different statistical approaches to analyze individual datasets from several INRA experiments. We implemented two strategies to analyze the continuum metadata. The first one consisted on beef qualities prediction using rearing practices. The second one consisted in achieving predefined objectives in terms of sensory quality, namely tenderness, using rearing practices in order to identify optimal management decisions for beef sector. The first example concerns an interesting statistical approach based on Principal Component Analysis (PCA) combined with the k-means clustering method after data standardization to create 3 groups of rearing practices. A dataset of 110 Rouge des Prés cows from different farms was used. For this, 16 rearing factors related to the animal's life and the finishing period were used to implement the ACP-k-means approach, which showed better results compared to PCA combined with hierarchical clustering analysis (HCA) and partitioning around medoids (PAM) to identify distinct groups of rearing practices. The second example explored the potential of chemometrics, based on partial least squares (PLS) combined with supervised learning methods (decision trees) for the identification of important variables from the continuum, using the first dataset of 110 Rouge des Prés cows to propose a prediction tool of tenderness. Overall, these statistical approaches showed the possibility to propose recommendations to take decision about the joint management of carcass and beef qualities to reach the targeted market specifications for both stakeholders and consumers.
Complete list of metadata

Cited literature [20 references]  Display  Hide  Download

https://hal.inrae.fr/hal-02738212
Contributor : Migration Prodinra <>
Submitted on : Tuesday, June 2, 2020 - 8:35:16 PM
Last modification on : Tuesday, April 13, 2021 - 3:34:21 AM
Long-term archiving on: : Wednesday, December 2, 2020 - 6:18:37 PM

File

Texte 3R_Gagaoua et al. 2018_1...
Publisher files allowed on an open archive

Identifiers

  • HAL Id : hal-02738212, version 1
  • PRODINRA : 455621

Collections

Citation

Mohammed Gagaoua, Brigitte Picard, Valérie Monteils. Étude du continuum pratiques d'élevage des bovins – carcasse – muscle – viande pour une gestion optimale des qualités des produits : stratégies statistiques applicables pour l’analyse de métadonnées. 24. Rencontres autour des recherches Ruminants (3R), Dec 2018, Paris, France. ⟨hal-02738212⟩

Share

Metrics

Record views

26

Files downloads

26