Toggle Main Menu Toggle Search

Open Access padlockePrints

Evaluation of a model to investigate the host-parasite interactions between first season grazing calves and O. ostertagi

Lookup NU author(s): Zoe Berk, Professor Ilias Kyriazakis


Full text for this publication is not currently held within this repository. Alternative links are provided below where available.


Implications We show that a simulation model predicting impacts of gastrointestinal parasitism on calves produces output results that agree well with independently produced experimental data, and we identify the most influential parameters. Introduction We have developed a model that predicts parasitological outputs (worm burdens and total egg outputs) and performance traits (food intakes and performance) for different infection modes and levels of infection in grazing calves (Berk et al., 2015). But before confidence can be placed in model predictions an evaluation must be conducted. The objectives of this study were to conduct a sensitivity analysis, and to compare model predictions with observations from relevant parasitological data, to validate the model concepts and parameters (Michel, 1970; Syzska and Kyriazakis, 2013). Material and methods A sensitivity analysis was conducted on model predictions for a subset of parameters using a Latin hypercube design. Parameter combinations covered a wide scope of possibilities, whilst having relatively few simulations (250). The mean parameter values were taken as the best estimate and all parameters were assumed to be normally distributed with CVs of 20%. To analyse the impact of variation in parameter on model predictions, an ANOVA was conducted for each parameter to test for significance (P <0.05). Independent sets of published experimental data were then used to independently validate model performance using both graphical comparisons and statistical tests of goodness-of-fit. Bias and root-mean-square error were calculated from differences in model predictions and reported data; these were tested for significance (P <0.05). Literature studies were selected as follows (1) Infections were O. ostertagi alone with no other parasite species involved; (2) all calves were infected during the growing phase; (3) calves were allowed access to adlibitum, high quality feed; (4) calves had no prior experience of parasitism at the start of the experiment. Comparisons reported here are for worm burdens (Michel, 1970) and faecal egg outputs (FECs) (Syzska and Kyriazakis, 2013). Results The sensitivity analysis revealed that model prediction of parasitological traits were most sensitive to the rate of immune development for host-controlled larval establishment and worm mortality; this implies the immune development rate has a large impact on worm burdens and subsequent parasitological outputs. Michel (1970) investigated worm burdens following trickle infections. The simulated and observed values followed the same pattern of increasing worm burdens up to a peak followed by a decrease: the correlation between predicted and observed values was r=0.78 (Figure 1). Although there was no systematic bias in the predicted results, not all predicted values fell within the one standard deviation (SD) of true values as indicated by the RMSE. Szyszka and Kyriazakis (2013) investigated the FEC for weekly larval challenge and found that FECs started low, increased to a peak approximately 35 d post infection and decreased thereafter. There was a positive correlation (0.75) with our model predictions (Figure 2). The predicted values were mostly within the one SD as indicated by the RMSE well within the 95% confidence interval; hence there was also no significant bias. Conclusion In general the developed model satisfactorily predicted parasitological traits, although after several comparisons with published literature (results not shown) the model was more effective at predicting the effects of subclinical rather than clinical challenges. Differences amongst independent published studies were large; this has been attributed to variability in calf genotypes and small samples sizes. The key question is whether development of a stochastic model incorporating variability between animals will account for the full range of experimentally observed outcomes. Acknowledgments The authors acknowledge funding from BBSRC and Merial References Berk, Z.L, Bishop, S, Forbes, A., Kyriazakis, I. 2015. Advances in Animal Biosciences 6 (2) 179 Michel, J.F. 1970. Parasitology 61, 435–447 Szyszka, O., Kyriazakis, I. 2013. Applied Animal Behaviour Science 147, 1–10

Publication metadata

Author(s): Berk Z, Bishop S, Forbes A, Kyriazakis I

Publication type: Conference Proceedings (inc. Abstract)

Publication status: Published

Conference Name: British Society of Animal Science Annual Conference - Science with Impact

Year of Conference: 2015

Pages: 180

Print publication date: 01/04/2015

Online publication date: 01/03/2015

ISSN: 2040-4700

Publisher: Cambridge University Press


DOI: 10.1017/S2040470015000035

Series Title: Advances in Animal Biosciences