Disease & heat resistance

Estimation of genetic variances and breeding values for infectious disease transmission host traits

---

Abstract

The transmission of infectious diseases within populations is regulated by three host traits: (i) susceptibility, being the propensity of an individual to become infected, (ii) infectivity, being the ability of a host to transmit the infection once infected and (iii) recoverability/mortality, determining the duration of the period during which the infected individual is infectious. Since these traits have shown to have individual variation and are partly genetically controlled, genetic selection on these traits could have substantial impact on reducing disease transmission and, thereby, the chances of epidemic outbreaks. However, to ensure a successful selection scheme, a suitable genetic evaluation method is needed to obtain reliable breeding value estimates (EBVs) for these traits. A method based on a generalized linear mixed model that allows to fit both susceptibility and infectivity simultaneously has been previously proposed. It has the theorical support within a quantitative genetic framework and it may be relatively easy to implement with common animal breeding software. However, studies testing the quality of the genetic variance and EBVs are still lacking. Here we used simulation to assess the accuracy of the estimates over a wide range of scenarios. A full/half-sib population of 2,400 individuals was created by mating 48 sires, each with 5 dams and producing 10 offspring per mating, resembling the structure of a pig population. Offspring were randomly divided into 240 independent groups of size 10, where an individual was infected and the epidemic allowed to progress. Latent phenotypes for susceptibility, infectivity and recoverability were simulated assuming the infinitesimal genetic model and a phenotypic variance of 0.25. The tested heritabilities for susceptibility and infectivity spanned from 0.1 to 0.8 (i.e. 64 scenarios, heritability for recoverability = 0.20) with at least 373 replicates in each scenario. The genetic variance estimates had large standard errors with a noticeable proportion of the replicates yielding zero estimates, especially for infectivity (i.e. 0.11 and 0.56 for susceptibility and infectivity, respectively), suggesting that larger populations are needed to obtain reliable estimates. The average accuracy of the susceptibility EBVs (across the heritabilities for replicates with non-zero genetic variance) was 0.33 (ranging from 0.18 to 0.42), and lower for infectivity averaging 0.13 (ranging from 0.06 to 0.19). As expected, the estimates (genetic variance and breeding values) improved with higher heritability of the trait in question, but they were little affected by the other trait heritability (i.e. susceptibility estimates were similar regardless of the infectivity heritability and the same for the infectivity estimates). Our results suggest that the method leads to good estimates for susceptibility but the genetic estimation of infectivity using GLMMs is more challenging. Alternative Bayesian inference approaches seem to improve the evaluation of infectivity, but their practical implementation needs to be validated.

Keywords: 2026

How to Cite:

Pong-Wong, R., Hulst, A., Prentice, J., Pooley, C., De Jong, M., Bijma, P. & Doeschl-Wilson, A., (2026) “Estimation of genetic variances and breeding values for infectious disease transmission host traits”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2285718. doi: https://doi.org/10.31274/wcgalp.23781

Rights: 1