This is a draft schedule. Presentation dates, times and locations may be subject to change.

117
Prediction Models for Standardized Total Tract Digestible Phosphorus in Swine Diets

Tuesday, July 11, 2017
Exhibit Hall (Baltimore Convention Center)
Jung Yeol Sung, Department of Animal Science and Technology, Konkuk University, Seoul, Korea, Republic of (South)
Beob Gyun Kim, Department of Animal Science and Technology, Konkuk University, Seoul, Korea, Republic of (South)
The objectives of the present study were to develop equations for predicting the standardized total tract digestible (STTD) P based on total P and phytate-P and supplemental phytase concentrations in swine diets and to evaluate the accuracy of the prediction models for STTD P using recently published data. A total of 414 data from 65 experiments that used supplemental phytase and determined P digestibility were employed to develop equations for predicting STTD P in swine diets. The CORR procedure of SAS was used to determine correlation coefficients among variables and the NLMIXED procedure was used to generate prediction equations. The STTD P was positively correlated with total P (r = 0.73; P < 0.001), phytate-P (r = 0.21; P < 0.001), and supplemental phytase (r = 0.36; P < 0.001) concentrations. Three equations for predicting the STTD P were: in the linear model, STTD P, g/kg as-fed = 0.244 + 0.660 × total P – 0.432 × phytate-P + 0.224 × phytate-P × phytase with Akaike information criterion (AIC) = 503.5 and Bayesian information criterion (BIC) = 523.6; in the quadratic model, STTD P, g/kg as-fed = 0.179 + 0.662 × total P – 0.422 × phytate-P + 0.286 × phytate-P × phytase – 0.011 × (phytate-P × phytase)2 with AIC = 498.8 and BIC = 523.0; in the exponential model, STTD P, g/kg as-fed = 0.214 + 0.664 × total P – 0.444 × phytate-P + 0.581 × phytate-P × (1 – e–0.603 × phytase) with AIC = 498.6 and BIC = 522.8, where total P range, 1.05 to 7.90 g/kg as-fed; phytate-P range, 0.40 to 6.18 g/kg as-fed; and phytase range, 0 to 2.22 FTU/g as-fed. According to the regression analysis for testing the accuracy of the developed quadratic model using recently published 42 STTD P data, the intercept (0.014 ± 0.041; P = 0.739) representing the mean bias and slope (−0.009 ± 0.051; P = 0.865) representing the linear bias were not different from 0. When validating the exponential model with the same method, the intercept (0.021 ± 0.038; P = 0.582) and slope (0.004 ± 0.048; P = 0.935) were also not different from 0. In conclusion, the equations suggested herein can fairly accurately predict STTD P concentrations using total P, phytate-P, and supplemental phytase concentrations as independent variables in swine diets.