Using near infrared spectroscopy of feces to predict growth performance in finishing feedlot cattle

Abstract

The objective of this thesis was to use fecal composition and apparent total tract digestibility (aTTD) as predicted by near infrared spectroscopy (NIRS) to assess growth performance in commercial feedlot cattle. Studies were conducted to 1) develop and validate NIRS calibrations using dried ground feces from cattle, 2) determine the optimal timing of sample collection, 3) compare samples collected from the pen floor versus rectum, 4) determine the optimal number of fecal samples to collect from a pen, 5) determine if NIRS could be used to detect changes in fecal nutrient concentrations and aTTD for a variety of diets, 6) find associations between fecal parameters and growth performance, 7) use NIRS of feed to predict energy content of grain screening pellets (GSP), 8) and assess the ability of fecal NIRS to predict dry matter intake (DMI), average daily gain (ADG), and gain to feed ratio (G:F) of feedlot cattle in a commercial feedlot. Fecal NIRS calibrations yielded accurate predictions (R2CV > 0.90, SECV < 2.42) for all fecal constituents except fat, and accuracy of predicting aTTD was high for starch (R2CV = 0.84, SECV = 1.06), moderate for DM, OM, CP, and GE (R2CV > 0.71, SECV < 2.88), but poor for NDF and ADF (R2CV < 0.33, SECV > 7.86). Most fecal nutrients and apparent total tract digestibility (aTTD) predictions varied over 24 h, however spot fecal samples collected at any time point from multiple cattle could be used as predictors of chemical composition and digestibility. Morning samples collected within 0 to 4 h after first feeding are optimal for estimating fecal starch and aTTD of starch. Except for DM, which was higher (P < 0.01) in pen floor than rectal fecal samples, there were minimal differences in fecal constituents between collection methods. When diets were fed containing wheat or increasing levels of silage in place of barley, aTTD of GE predicted using NIRS was related to net energy of gain (NEg) of the diets as estimated by performance (R2 = 0.58, P = 0.03 and R2 = 0.43, P < 0.01, respectively). Similarly, observed ADG could be predicted using NIRS of feces for steers fed wheat (R2 = 0.48, P = 0.05) and increasing levels of silage (R2 = 0.40, P < 0.01), but not G:F. Compared to measured performance data, NIRS over predicted the energy content of grain screening pellets. When comparing cattle of different sexes, different processing methods and grain types, a quadratic relationship was observed between fecal starch, sex, average BW at time of sampling, and G:F (rho = 0.75, P < 0.01). These data indicate that NIRS predictions using dried ground feces collected from the pen floor from multiple feedlot cattle could predict G:F of feedlot cattle with reasonable certitude when variables such as BW and sex were included in the model.