Trait deviations were corrected for herd year season effects, permanent environmental effects, and heterosis.
The phenotype considered was LEPG, corrected for herd effect, parity effect and effect of days in milk as estimated from the complete data set (4,053 cows).
The model corrected for herd classification (i.e. high-incidence and low-incidence herds), trimming season, breed, parity, lactation stage, and interactions of herd classification and trimming season and herd classification and parity (same effects as in the final model as explained later), and included herd and cow as random effects.
To allow a comparison between herds of different sizes this final Appleby (Appleby, 1980) rank score (a) was adjusted to correct for herd size (h), using: adjusted dominance rank = 1 − a − 1 h − 1 This 'adjusted dominance rank' was the outcome variable used in the statistical analysis, referred to as dominance rank.
Phenotypes of bulls and cows were constructed as daughter trait deviations (the average of the bull's daughters trait deviations corrected for breed of mate) and trait deviations, respectively (corrected for herd year season and permanent environment effects) [see Additional file 2: Table S2].
Briefly, the model fitted in both BayesA and BAYES_SSVS was: where y is a vector of n daughter yield deviations corrected for herd year season effects for each trait, X is (n × m) a design matrix allocating records to the marker effects with element X ij = 0, 1 or 2 if the genotype of animal i at SNP j is 11, 12 or 22 respectively.
To correct for herds with multiple test observations over the 8 year period, the farm CPH code was included as a random effect in the model.
Thus applying the estimates in the sample size calculation formula for simple random sampling, and correcting for a finite population we planned to sample 125 herds represented as 53, 48 and 24 herds for Blue Lagoon, Lochinvar and Kazungula, respectively.
Where the true number of infected herds was obtained by adding the false negative and subtracting the false positive while correcting for the specificity and sensitivity of the test previously estimated to 0.89 and 0.88 respectively [ 12].
