Predicting beef palatability using conventional and rapid analytical methods: A random forest approach
Abstract
This study evaluated the ability of traditional and rapid analytical methods to predict consumer beef liking using random forest modelling. Eighty beef samples from four muscles and four animal classes were analysed for intramuscular fat and fatty acid composition (IMF-FA), other meat quality traits (MQ: pH, colour, shear force, cook loss), near-infrared spectroscopy (NIRS), and rapid evaporative ionisation mass spectrometry (REIMS). Consumer sensory evaluation ( n = 180) followed Meat Standards Australia protocols, with liking quantified using the composite MQ4 score. Models were developed using these individual datasets and their combinations and evaluated on independent test data. MQ traits provided the highest predictive accuracy of MQ4 (R 2 = 0.62, RMSE = 7.31), indicating that key meat properties capture most variation in consumer liking. IMF-FA showed moderate performance (R 2 = 0.43, RMSE = 8.62), reflecting that lipid variables contained some predictive information for MQ4. REIMS achieved R 2 = 0.55 and RMSE = 8.99 using metabolite and lipid features tentatively associated with post-mortem biochemical processes and flavour precursors. NIRS showed limited predictive performance under the conditions evaluated (R 2 = 0.17, RMSE = 10.27). No combined model exceeded MQ alone. Relative to IMF-FA alone, adding MQ improved all performance metrics, adding REIMS to MQ + IMF-FA increased R 2 but not RMSE or MAE. Among the methods evaluated, conventional MQ traits provided the highest predictive performance for consumer palatability. REIMS showed lower predictive ability than MQ traits but warrants further evaluation as a rapid tool for predicting consumer liking. These findings reflect the relative performance of the datasets for predicting MQ4 within the population and experimental conditions evaluated and require validation using larger independent datasets.
Related
- Prevalence and quantification of viable Campylobacter species in retail chicken: a comparative study using viability digital polymerase chain reaction and conventional culture methods
- Review of Quality Attributes and Analytical Methods Used for Comparative Analytical Assessment of Monoclonal Antibodies as Part of Successful Biosimilar Approvals in the United States and European Union
- Analytical Methods for Fluid Biomarkers in Alzheimer’s Disease from Discovery to Clinical Implementation
- Toxin structure shapes palatability in a chemically defended butterfly
- Comparative palatability and efficacy of two praziquantel isomers in rainbow trout infected by eyefluke.
BibTeX
@article{realini2026predicting,
title = {{Predicting beef palatability using conventional and rapid analytical methods: A random forest approach}},
author = {Carolina Realini and Renyu Zhang and Dongwen Luo and Marlon Reis and Yash Dixit and Alastair Ross and Vanessa Ruppert and Jean-Francois Hocquette and Alix Neveu and Rod Polkinghorne and Marie-Pierre Ellies-Oury},
journal = {Meat Science},
year = {2026},
month = oct,
doi = {10.1016/j.meatsci.2026.110231},
url = {https://hal.science/hal-05772900v1},
note = {CC BY},
}
RIS
TY - JOUR
TI - Predicting beef palatability using conventional and rapid analytical methods: A random forest approach
AU - Carolina Realini
AU - Renyu Zhang
AU - Dongwen Luo
AU - Marlon Reis
AU - Yash Dixit
AU - Alastair Ross
AU - Vanessa Ruppert
AU - Jean-Francois Hocquette
AU - Alix Neveu
AU - Rod Polkinghorne
AU - Marie-Pierre Ellies-Oury
PY - 2026
DA - 2026/10/01
JO - Meat Science
DO - 10.1016/j.meatsci.2026.110231
UR - https://hal.science/hal-05772900v1
AB - This study evaluated the ability of traditional and rapid analytical methods to predict consumer beef liking using random forest modelling. Eighty beef samples from four muscles and four animal classes were analysed for intramuscular fat and fatty acid composition (IMF-FA), other meat quality traits (MQ: pH, colour, shear force, cook loss), near-infrared spectroscopy (NIRS), and rapid evaporative ionisation mass spectrometry (REIMS). Consumer sensory evaluation ( n = 180) followed Meat Standards Australia protocols, with liking quantified using the composite MQ4 score. Models were developed using these individual datasets and their combinations and evaluated on independent test data. MQ traits provided the highest predictive accuracy of MQ4 (R 2 = 0.62, RMSE = 7.31), indicating that key meat properties capture most variation in consumer liking. IMF-FA showed moderate performance (R 2 = 0.43, RMSE = 8.62), reflecting that lipid variables contained some predictive information for MQ4. REIMS achieved R 2 = 0.55 and RMSE = 8.99 using metabolite and lipid features tentatively associated with post-mortem biochemical processes and flavour precursors. NIRS showed limited predictive performance under the conditions evaluated (R 2 = 0.17, RMSE = 10.27). No combined model exceeded MQ alone. Relative to IMF-FA alone, adding MQ improved all performance metrics, adding REIMS to MQ + IMF-FA increased R 2 but not RMSE or MAE. Among the methods evaluated, conventional MQ traits provided the highest predictive performance for consumer palatability. REIMS showed lower predictive ability than MQ traits but warrants further evaluation as a rapid tool for predicting consumer liking. These findings reflect the relative performance of the datasets for predicting MQ4 within the population and experimental conditions evaluated and require validation using larger independent datasets.
SN - 0309-1740
SN - 0309-1740
ER -