Some abstracts do not have video files because ASAS was denied recording rights.

1494
Effects of animal and diet characteristics on digestibilities of dry matter, fiber and starch in lactating cows

Friday, July 22, 2016: 3:30 PM
155 E (Salt Palace Convention Center)
Rodrigo A De Souza , Michigan State University, East Lansing, MI
Robert J Tempelman , Michigan State University, East Lansing, MI
Mike S Allen , Michigan State University, East Lansing, MI
John K Bernard , University of Georgia, Tifton, GA
Bill Weiss , The Ohio State University, Wooster, OH
Mike J VandeHaar , Michigan State University, East Lansing, MI
Abstract Text: Our objective was to determine the effects of DMI, BW, and diet characteristics on total tract digestibilities of DM, NDF and starch (DMD, NDFD and StarchD, respectively) in high producing cows. The database was constructed with individual observations of digestibilities, ingredients and chemical composition of diets, BW and DMI, experimental design, and treatment arrangement.  The dataset contained 1,942 observations on 635 cows from 56 studies with >6 observations per treatment in MI, OH, and GA. Forage sources included corn, alfalfa, wheat straw and orchardgrass. Nutrient digestibility means were 66 ± 6, 42 ± 11 and 93 ± 5% for DMD, NDFD and StarchD, respectively. Diet nutrient contents (%DM) were 31 ± 5% NDF, 27 ± 6% Starch, 2.6 ± 1.2% FA, 17 ± 1.4% CP. DMI and BW were 23 ± 5 and 669 ±79 kg, respectively. Data was analyzed using a mixed model with HPMIXED procedure of SAS. The full model included linear and quadratic effects of diets, BW0.75, DMI, and all possible 2-way interactions between diet variables and DMI as fixed effects. Cow, block, period and study were used as random effects. Best fitting reduced models were generated using backward and stepwise regression procedures. Interactions were not significant. A simplest model was generated using only DMI and the random effects. The backward, stepwise, and simplest models were cross-validated using five folds, and resulting correlations coefficients across studies (CORR) were compared by t-test. The model that resulted in the highest CORR and lowest number of variables was accepted as the best fitting model for each nutrient digestibility. The best fitting prediction equations were: DMD = 69 – 0.20 * DMI (MSE=8.5, r2=0.68, CORR=0.20); NDFD = 32 + 0.26 * grass + 0.46 * starch + 3.4 * FA + 0.83 * DMI –0.020 * starch2 – 0.34 * FA2 – 0.030 * DMI2 (MSE=26, r2=0.80, CORR=0.41); and StarchD = 97 – 0.30 * DMI (MSE=3.1, r2=0.84, CORR=0.22); grass, starch and FA are expressed as percentage of DM and DMI is expressed as percentage of BW0.75. Our results confirm that digestibility was reduced as DMI increased, but at a lower rate than the previous equation used by NRC (2001).  DMD and StarchD can be predicted based on only DMI; however, NDFD required diet characteristics in addition to DMI. Starch levels >25% and FA >5% dramatically reduced NDFD.

Keywords: Digestibility, Model, Intake, NDF, Starch