Prediction of rumen volatile fatty acids based on milk fatty acid profiles

The objectives of this study were to develop and validate models for predicting rumen volatile fatty acid proportions of dairy cows based on milk fatty acids. Model development, using data from two experiments, showed predictive equations explained 30 to 87 % of the variation in volatile fatty acids. Validation of the equations, using an independent experiment, showed predictions of acetate and propionate proportions were relatively accurate, with a prediction error lower than 10 % of the observed mean.