The NS Dairy Cattle : protein requirements submodel for lactation dairy cows

Abstract

Protein is one of the most costly nutrients in the diet, and its use in inadequate proportions can be cause productive and economics losses and environmental problems. Therefore, our objective was to develop a new protein requirement system for maintenance and lactation for dairy cows. A large dataset, containing 1042 experimental diets from 269 experiments published in 260 peerreview papers was built. The data set was divided into two independent subsets. The first data subset contained only nitrogen (N) balance trials and it was used to develop equations to protein requirements for maintenance and lactation (212 peer-review papers; 221 experiments; 856 experimental diets). The second data subset was used to assess the adequacy of the predictive models (48 peer-review papers; 48 experiments; 182 experimental diets), and it contained only experiments that measured non ammonia non microbial N (NANMN) and microbial N intestinal fluxes were measured in lactating dairy cows. The requirement of net protein for maintenance (NPM) was calculated as intercept of the nonlinear meta-regression between N intake and N fecal + urinary excretion (NFU). Feed system (TMR or pasture; P = 0.15), genetic group (Bos taurus vs Bos taurus × Bos indicus; P = 0.16), lactation stage (< 100 vs ≥ 100 days in milk (DIM); P = 0.55) fecal output method (total collect vs indirect; P = 0.88) and urinary output method (total collect vs spot; P = 0.61) did not affect NPM. Therefore, we proposed an overall value to NPM (6.52 ± 0.15 g protein/kg BW0.75). Feed system and days in milk affected (P ≤ 0.05) the efficiency of conversion of metabolizable protein (MP) to maintenance net protein (EMPM); EMPM (0 to 1): [0.8287 + feed system (TMR = 0, Pasture = - 0.0441) + DIM (<100 = 0, ≥ 100 = 0.1239)]. We estimated efficiency of use of MP for lactation (EMPL) fixed and variable. Genetic group affected the (P ≤ 0.05) the efficiency of conversion of MP to lactation (EMPL). Therefore, we proposed used of fixed value to EMPL according with genetic group (Bos taurus = 0.71 ± 0.02 and Bos taurus x Bos indicus crossbred = 0.55 ± 0.09), or EMPL variable according with N milk (g N milk/kg dry mater intake, DMI) = [0.4493exp0.0684×Nmilk (g/kgDMI)]. The proposed models NS Dairy Cattle I (EMPL Fixed) e II (EMPL Variable) presented similar values (P > 0.05) of root mean square prediction error (RMSPE; 19.35 and 20.05% MP observed) and concordance correlation coefficient (CCC; 0.57 and 0.50), indicating similar accuracy and precision to predict MP. To compare with extent models, our models predict MP with lower RMSPE (P < 0.05) than INRA (2007), CSIRO (2007), AFRC (1993) and DVE (2007) models, but it was similar (P > 0.05) to NRC (2001), Feed into Milk (FIM, 2004) and CNCPS (6.5, 2015) models to predict MP in lactating dairy cows. Because it is the only MP model built with lactating dairy cows and because of its good accuracy and precision, we recommend the use of the NS Dairy Cattle model to predict MP for maintenance and lactation of dairy cows.