Phenomics

Integration of near-infrared spectra in the genetic evaluation of meat quality traits in Duroc pigs

---

Abstract

Near-infrared (NIR) spectra are traditionally used as predictor for compositional traits, but the raw information from the NIR spectra can also be used for increasing the accuracy of genetic evaluations. NIR spectra is advantageous compared to other omics data because it can be obtained at the large scales that are required for genetic evaluations in commercial breeding programs at virtually no cost and using non-destructive methods. The objective of this study was to evaluate the use of NIR spectra for genetic evaluation of carcass and meat quality traits in pigs. We used data from a commercial purebred Duroc line in which NIR spectra (2,001 wavelengths: from 500 to 2,500 nm) have been recorded for more than 17,000 pigs over almost a decade, from the loin exposed surface at the cutting plant. Carcass weight and backfat thickness were recorded at slaughter. Pork pH was determined in the semimembranosus muscle and intramuscular fat content (IMF) and fatty acid composition were determined in the gluteus medius muscle by gas chromatography in a subset of approximately 2,300 pigs. Genetic evaluations were performed either on univariate models for each of the carcass (weight and backfat thickness) and meat quality (pH, IMF, and total contents of saturated, monounsaturated, and polyunsaturated fatty acids) traits or on multi-trait models. In the multi-trait models, the first trait was each of the carcass and meat quality traits while the secondary traits were the two first principal components (PC) derived from the whole preprocessed NIR spectra or the first PC of each of 100-nm bins of the spectra. Prediction accuracy was tested by masking the phenotypic data from the most recent batches. The heritability of the two first PC of the preprocessed spectra were quite low despite explaining most of the variance of the NIR spectra (PC1: 26% of variance, h2=0.15; PC2: 13% of variance, h2=0.07) and that of the PC1 of each of the 100-nm bins (9 to 99% of variance) ranged from 0.01 to 0.24. Genetic correlations reached 0.19 for carcass weight, 0.27 for backfat thickness, 0.59 for pH, 0.70 for IMF, and 0.33 to 0.97 for fatty acid composition. Compared to the univariate models, prediction accuracies increased when some PC of the NIR spectra were added as a secondary trait, especially for the meat quality traits. These results support the potential of phenomic approaches based on NIR spectra from fresh pork for pig breeding programs. Such potential, either using the NIR spectra as secondary traits or for spectra-based relationship matrices, is currently being benchmarked against standard genomic prediction models.

Keywords: 2026

How to Cite:

Estany, J., Gol, S., Reixach, J. & Ros-Freixedes, R., (2026) “Integration of near-infrared spectra in the genetic evaluation of meat quality traits in Duroc pigs”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286394. doi: https://doi.org/10.31274/wcgalp.23979

Rights: 1