Skip to main content
ARS Home » Plains Area » Miles City, Montana » Livestock and Range Research Laboratory » Research » Publications at this Location » Publication #418441
Research Project: Identifying and Mitigating Factors that Limit Beef Production Efficiency

Location: Livestock and Range Research Laboratory

Title: Machine learning for genomic prediction of growth traits in a composite beef cattle population

Author
item Hay, El Hamidi

Submitted to: Animals
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/17/2024
Publication Date: 11/18/2024
Citation: Hay, E.A. 2024. Machine learning for genomic prediction of growth traits in a composite beef cattle population. Animals. 14(20). Article 3014. https://doi.org/10.3390/ani14203014.
DOI: https://doi.org/10.3390/ani14203014

Interpretive Summary: Genomic selection is commonly used in many livestock species to predict important traits. However, the current methods used to make these predictions aren't perfect. New approaches, like machine learning, could make these predictions more accurate because they can handle complex relationships in the data. In this study, we tested two machine learning methods, Random Forest and Support Vector Machine, to predict birth weight, weaning weight, and yearling weight in a beef cattle population. We compared these methods with two traditional ones, GBLUP and BayesA. The GBLUP method was the most accurate for predicting birth and yearling weights, while the Random Forest method was better at predicting weaning weight. Additionally, GBLUP provided a closer match to the actual data. Overall, GBLUP gave better predictions and fit the data better than the machine learning methods we tested.

Technical Abstract: The adoption of genomic selection is prevalent across various plant and livestock species, yet existing models for predicting genomic breeding values often remain suboptimal. Machine learning models present a promising avenue to enhance prediction accuracy due to their ability to accommodate both linear and non-linear relationships. In this study, we evaluated two machine learning models, Random Forest and Support Vector Machine, for predicting genomic values related to birth weight (BW), weaning weight (WW), and yearling weight (YW), and compared them with two conventional models, GBLUP (Genomic Best Linear Unbiased Prediction) and BayesA. The results demonstrated that the GBLUP model achieved the highest prediction accuracy for both BW and YW, whereas the Random Forest model exhibited superior prediction accuracy for WW. Furthermore, GBLUP outperformed the other models in terms of model fit, as evidenced by lower mean square error values and regression coefficients of corrected phenotypes on predicted values. Overall, the GBLUP model delivered superior prediction accuracy and model fit compared to the machine learning models tested.