Process & Systems Modelling

Bayesian inference of genetic and epidemiological
parameters from field disease data

Field level disease incidence data enable quantification of host genetic variation for disease resistance and are invaluable in identifying appropriate breeding strategies aimed at enhancing genetic resistance to disease. However current methods of analysis do not account for the often incomplete (e.g. due to lack of information on exposure to disease) and noisy (due to imperfect diagnostic tests) character of such data, and this can lead to underestimation of the true extent of genetic resistance.

A Bayesian inferential framework has been developed to quantify host genetic variation in a manner which accounts for the complexities inherent in field disease data. The framework integrates genetic and epidemiological concepts with field disease data incorporating genetic relationships between animals, observed disease state, relative prevalence of the disease and sensitivity and specificity of the diagnostic test. Using the simulated data, we have found that the framework enables inference on both genetic (e.g. heritability of resistance to disease) and epidemiological (e.g. prevalence of disease) parameters that are of practical relevance to animal breeders.

inderred probability surface Inferred probability surface for heritability of susceptibility to a disease (x-axis) and sensitivity of the diagnostic test (y-axis) are plotted, indicating accurate inference for a simulated data set with heritability of 0.25 and test sensitivity of 0.85.

Further details from: Mintu Nath

Article date 2013

Research

Statistical Genomics and Bioinformatics

Process and Systems Modelling

Statistical Methodology

PhD Opportunities

Meetings & Seminars