Disease & heat resistance

Impact of phenotypic pre-adjustments on genetic parameters for disease resilience traits in pigs

---

Abstract

Genetic selection for enhanced disease resilience in pigs requires phenotypic records, genomic data, and statistical models that accurately capture meaningful biological variation under infection. We quantified the potential impact of trait definition (i.e., continuous vs categorical) and model adjustments for body weight and/or age at challenge on genetic parameter estimates, prediction accuracy, and potential re-ranking of individuals for disease resilience traits. Data for 5,395 crossbred pigs from three wean-to-finish artificial challenge trials, conducted between 2018 and 2023, were available for analyses. Average daily gain (ADG) was computed as change in body weight (BW) per day using all available data recorded before (ADGpre), during (ADGdur), and after (ADGpos) inoculation, and from wean-to-market (ADGall). Additional traits included: clinical score at 0, 13, 21, and 42 days post-inoculation (dpi) and the sum of clinical scores across days (CumCS); days-to-death or market (DTDM) and its inverse normal transformation (DTDIN); mortality from 0 to 28 dpi (MOR28) and from wean-to-market (MORall); total treatments administered post-challenge (TRTc), as well as treated at least once (TRTb). Each trait was analyzed twice, using identical models, except that for Model 1, only age at challenge was fitted as a covariate and for Model 2, both age and BW at challenge were fitted as covariates. Genetic parameters, prediction accuracies, and genomic estimated breeding value (gEBV) rankings were compared between models using Pearson and Spearman correlations and proportion of commonly selected animals (accuracy ≥ 0.45). Heritability of CumCs, DTDM, DTDIN, and TRTc were 0.17, 0.11, 0.14, and 0.10, respectively. Corresponding genetic correlations with their categorical counterparts were: 0.49, 0.99, and 0.99 for CumCS with CS at 0, 13, and 42 dpi, respectively; -0.81 and -0.78 for DTDM with MORall and MOR28, respectively; -0.26 and -0.35 for DTDIN with MOR28 and MORall, respectively; and 0.98 for TRTc with TRTb. Therefore, CumCS, DTDM, and TRTc were highly correlated with their categorical counterparts, indicating they can serve as effective auxiliary resilience traits to support selection decisions. Adjusting for BW at challenge (Model 2) reduced additive genetic variance and heritabilities for ADG traits, and on average, lowered prediction accuracy for ADG and clinical/treatment traits (mean difference = -0.01 to -0.12), while improving accuracies for survival traits (mean difference = 0.00 to +0.03). Despite shifts in variance partitioning, cross-model gEBV correlations were high ( >0.80) and clear re-ranking was uncommon, with a notable exception for ADGpre. Traits recorded post-challenge were robust to model adjustments. Lastly, genetic correlations between ADGdur and ADGpos indicate that greater ADG under disease pressure was associated with reduced clinical signs, treatments, and mortality rate post-challenge, indicating that any of these traits could be used to select for enhanced resilience to disease challenge.

Keywords: 2026

How to Cite:

Souza, M., Dunkelberger, J., Sevillano, C. & Brito, L., (2026) “Impact of phenotypic pre-adjustments on genetic parameters for disease resilience traits in pigs”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2286516. doi: https://doi.org/10.31274/wcgalp.24036

Rights: 1