Disease & heat resistance

From hidden transmission of MAP to heritable susceptibility: genetic parameter estimation using simulated data

---

Abstract

Paratuberculosis (Johne's disease), caused by Mycobacterium avium subspecies paratuberculosis, remains a persistent challenge in dairy production despite decades of research and control programs. Selective breeding for reduced susceptibility has been proposed as a strategy to achieve disease eradication in previous modelling studies. However, accurately estimating the genetic parameters of susceptibility and determining how closely they reflect the true underlying variance remains challenging because the disease transmission process, in particular the exposure to infection, cannot be directly observed in field data. Simulation offers a powerful approach to develop methods to address these challenges. In this study, we developed an agent-based genetic-epidemiological model to simulate MAP transmission under a realistic dairy herd setup, linking host genetics to infection dynamics. Individual susceptibility was determined by a polygenic architecture consisting of 1,000 additive loci distributed across the bovine genome. Locus effects were sampled from a normal distribution, resulting in a genetic variance of 0.25 for log-susceptibility, corresponding to an observable heritability of approximately 0.03. Herds of 200 animals were simulated over five years, under a demographic structure representative of commercial dairy systems. Infection events included both vertical (dam-to-calf) and environmental transmission within the calving, calf, and adult pens. Using these infection histories, we then simulated field data reflecting what could be observed under practical conditions. Testing scenarios differed in frequency (annual, seasonal, monthly), age groups tested (all animals or only adults over two years), and result type (binary or categorical). These scenarios represent both ideal and realistic surveillance conditions, producing datasets with varying diagnostic accuracy and completeness. Genetic parameters and breeding values were estimated using a generalized linear mixed model (GLMM) based on the transmission dynamics, with a complementary log-log link and exposure offset. We further compare it with a linear mixed model (LMM) and an extended LMM incorporating infection exposure as a covariate. In results, simulated populations with genetic variation in susceptibility reached an endemic prevalence of 47.7%, with most new infections occurring in calves (81.2%), followed by newborns (15.7%) and adults (3.03%). Under the baseline GLMM with annual testing and binary outcomes across all age groups (approximately 6,131 records), the estimated genetic variance of susceptibility was 0.20, with cow- and sire-level breeding value accuracies of 0.25 and 0.35, respectively (on average 20 offspring per sire). In contrast, the LMM and extended LMM estimate genetic variance on the binary outcome scale, leading to substantial deviation from the susceptibility variance. More results on accuracies and alternative scenarios will be presented at the conference. Overall, the main benefit of the GLMM is that it provides epidemiologically interpretable estimates of genetic variance and breeding values for susceptibility, making them meaningful for predicting response to selection.

Keywords: 2026

How to Cite:

Gao, Y., Hartemink, N., de Jong, M. & Bijma, P., (2026) “From hidden transmission of MAP to heritable susceptibility: genetic parameter estimation using simulated data”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2266828. doi: https://doi.org/10.31274/wcgalp.23396

Rights: 1