Evaluating Computer-Based Rumen Protein Models

As dairy consultants and managers use computer-based rumen models to fine-tune feeding programs, questions are raised. How effective are these programs? What information do they provide? Do software programs provide similar results? This study evaluated protein relationships and results.

A standard ration was balanced using a university software program (Spartan II, Michigan State University) that did not have a rumen modeling software. The ration consisted of 20 pounds of corn silage, 10 pounds of alfalfa haylage, 9 pounds ground shelled corn, 4 pounds of 44 percent soybean meal, 2 pounds of heat treated soybean meal, 3 pounds of corn distillers grain, 0.50 pound of blood meal, 0.20 pound of urea, 3 pounds of fuzzy cottonseed, and 1 pound of soy hulls (all feed amounts are on a 100% dry matter basis). Cow specifications included 1350 pounds of body weight, 100 pounds of milk, 3.5 percent milk fat, 3.3 percent total protein, mature cow, and not pregnant. No minerals, vitamins, or additives were included. Feed ingredients and amounts were selected to meet Spartan requirements using typical Illinois ingredients fed at typical levels. The same cow specifications and feed qualities were entered in Dairy NRC 2001, Cornell Net Carbohydrate Protein System (CNCPS) version 4.10.13, and AminoCow (DeGussa) models.

The three models had similar levels of dry matter intake, crude protein, NDF, NFC, and fat. Differences in metabolizable protein (MP) reflect variation in model equations and feed library values (all feed crude protein values were standardized for each model). All three rumen programs indicated that excess rumen degraded protein (RDP) was fed while rumen undegraded protein (RUP) was low. Each program provided values that were useful (potential milk, soluble protein, rumen pH, and amino acid requirements). Amino acid levels and MP levels varied between models. Dairy nutritionist should consider using a rumen modeling program to fine-tune rations realizing each program will provide different features and results. Select a program that is easy to run and economical to purchase; then monitor farm results. All three programs performed satisfactorily.