The purpose of this research was to evaluate four objectives: 1) update and evaluate predictions of essential amino acid (EAA) outflows from the rumen, 2) predict EAA use and release by the portal drained viscera (PDV) and liver (LIV) of dairy cows, 3) predict EAA use by the mammary (MAM) and non-splanchnic, non-mammary (OTH) tissues, and 4) predict milk protein production from MAM use. To evaluate the first objective, a model was constructed using previously derived equations for ruminally undegraded (RUP), microbial (MiP) and endogenous protein (EndP) flow from the rumen and refit to literature data. Corrections were included in the model to address recovery of EAA during 24-h acid hydrolysis. Upon initial evaluation, all EAA, except Leu, were over predicted and slope bias (P < 0.01) was present for all except Met and Leu. Because of the bias, residuals were regressed on the EAA from each protein flow and adjustments were made to the protein flows. The added adjustments removed all mean bias for the EAA; however, a small slope bias was introduced for Lys and Thr. To evaluate the second objective, equations of Hanigan et al. (2004b) were tested and modifications were made to determine which equation form best represented EAA use by the tissue. Upon initial evaluation of the PDV model of Hanigan et al. (2004b), significant slope bias was present and addressed by deriving alternative forms of the equation. Initial predicted EAA use displayed a mean bias ranging from 0.15 to 45 % and a slope bias ranging from 0.02 to 76% mean square error. The alternative equation forms derived reduced the overall mean and slope bias and improved other fit statistics (RMSE, CCC). To evaluate the third objective, previously derived equations from Hanigan et al. (1998b) were tested using literature data and modifications were made to address deficiencies for each EAA. Upon initial evaluation of the MAM model, significant mean and slope bias was present and was further addressed by derivation of alternative equation forms. Initial evaluation of the OTH model displayed significant mean and slope bias for majority of the EAA ranging from 0.3 to 26 % for mean and 46 to 61 % for slope. For the last objective, several models, both linear and non-linear were evaluated to determine which EAA have a significant impact on milk protein. All models derived has prediction errors below 18-20 % which is comparable or a s light improvement as compared to previous literature data (Moraes et al., 2018). Overall, the equations evaluated show promise in accurately predicting dietary EAA from the time of absorption to their use within the tissues (PDV, LIV, MAM, and OTH) and further impact on milk protein production. / The purpose of this research was to develop a system to predict the supply of essential amino acids (EAA), tissue use of the absorbed EAA, and milk protein production in dairy cows. Essential amino acid supply was predicted from previously published predictions of protein flows and revised estimates of amino acid composition of the respective proteins with corrections for amino acid recovery from acid hydrolysis. Use was predicted for the portal drained viscera (PDV), liver (LIV), mammary (MAM), and an aggregation of non-splanchnic, non-mammary tissues (OTH). Splanchnic EAA release (supply – use) was predicted with root mean squared errors of the mean observed that were less than 12%, and concordance correlation coefficients greater than 0.85 with mean and slope biases ranging, respectively, from 0 to 76% and 0 to 45% of mean squared errors. Mammary tissue EAA use had root mean squared errors ranging from 28 to 148% of the observed mean and concordance correlation coefficients ranging from -0.04 to 0.48. Arterial EAA concentrations and use by OTH was predicted as components within the overall system. OTH utilization was predicted with root mean squared errors < 6% of the observed mean and concordance correlations near 1.0. Alternative equation forms were also derived for each tissue that resulted in marginal reductions in mean and slope biases. Milk protein production was predicted from predicted mammary EAA utilization and digestible energy supply with a root mean squared error of less than 18% which is an improvement over existing field application models. Based on model performance, it was concluded that the system could be used in field application software which would allow more accurate ration formulation resulting in reduced nitrogen excretion by dairy cattle.
| Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/95890 |
| Date | 08 June 2018 |
| Creators | Myers, Adelyn |
| Contributors | Dairy Science |
| Publisher | Virginia Tech |
| Source Sets | Virginia Tech Theses and Dissertation |
| Detected Language | English |
| Type | Thesis |
| Format | ETD, application/pdf |
| Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
Page generated in 0.0019 seconds