Skip to main content
ARS Home » Plains Area » Woodward, Oklahoma » Rangeland and Pasture Research » Research » Publications at this Location » Publication #368820

Research Project: Sustaining Southern Plains Landscapes through Improved Plant Genetics and Sound Forage-Livestock Production Systems

Location: Rangeland and Pasture Research

Title: Comparing two software programs for fitting nonlinear, one- and two-compartment age-dependent digestion models: empirical data

Author
item Gunter, Stacey
item GADBERRY, SHANE - University Of Arkansas
item COFFEY, KEN - University Of Arkansas
item Moffet, Corey

Submitted to: Animal Production Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/1/2022
Publication Date: 7/15/2022
Citation: Gunter, S.A., Gadberry, S., Coffey, K., Moffet, C. 2022. Comparing two software programs for fitting nonlinear, one- and two-compartment age-dependent digestion models: empirical data. Animal Production Science. https://doi.org/10.1071/AN21311.
DOI: https://doi.org/10.1071/AN21311

Interpretive Summary: To be able to predict the performance of grazing cattle in extensive rangeland environments, herbage intake is paramount because it quantifies energy intake and performance. Digestion models have been developed to predict the movement and residence time of feedstuffs with in the digestive systems of cattle. Many of the digestion models develop more than 20 years ago were developed to be parameterized on specific statistical software systems. Recently, new statistical software systems have become available and are being used by researchers. These new systems need to be tested compared to older systems to validate that parameters produced by either systems are comparable. In the experiment, we compared the newer R software (The R Foundation for Statistical Computing; Vienna, Austria) to the established SAS software (SAS Inst., Inc.; Cary, NC) using the nonlinear regression function to parameterize a one-compartment, age-dependent models. We found that both software systems produced similar parameters and digesta kinetics estimates by either software system were similar, and biases were small. Based on this analysis, the R software is acceptable to fit one- and two-compartment, age-dependent models by nonlinear regression compared to SAS software.

Technical Abstract: With a need for greater efficiency in ruminant production, using compartmental modeling to access the quantity and extent of digestion within the gastrointestinal tract is a valuable tool. To improve the utilization of these models, we examined the biases and SD of the one- (G2) and two-compartmental (G2G1) models with a gamma-2 distribution in the age-dependent compartment when parameterized with two different software programs using 41 datasets of Yb concentrations in fecal samples collected at discrete times after a pulse-dose of Yb was administered obtained from published research. The resulting fecal marker concentration datasets were fit to G2 and G2G1 models with programs written for 2 software systems (R and SAS). The resulting model parameters, K0, ' or '1, K2, and t, were used to calculate the digesta kinetics parameters: particle passage rate, gastrointestinal DM fill, fecal DM output, gastrointestinal mean retention time, and rumen retention time. We evaluated the bias and SD of the parameters and digesta kinetics across 3 percentile groups (20, 50, and 80 ± 15%) and between software programs. When datasets were fit to the G2 model, all converged for both software programs. But, when fitting the same datasets to the G2G1 model, a small number did not converge with the SAS program (n = 3). Bias and SD of differences between software packages were small, however G2G1 model produced smaller bias and SD of differences than G2 model. Bias and SD for digestion parameters across the percentile groups did not vary linearly for most parameters and digesta kinetics and the biases and SD of differences were small within groups. Hence, model parameters and digesta kinetics estimates from R or SAS software programs can be used interchangeably when using these parameters and estimates in nutritional modelling.