Fecal NIRS prediction of dietary protein percentage and in vitro dry matter digestibility in diets ingested by goats in Mediterranean scrubland
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
A method to elucidate the nutritional value of goats' diets in ligneous environments is needed. Near infrared spectroscopy (NIRS) can be applied to fecal material to establish statistical relationships between the reflectance of fecal samples in the near infrared region and the diets consumed. We tested fecal NIRS for the prediction of dietary crude protein content (CP) and in vitro dry matter digestibility (DMD) in Damascus goats browsing Mediterranean scrubland dominated by Pistacia lentiscus and Phyllirea latifolia. Two datasets consisting of pairs of feces and dietary data were used; for the first (n = 151), goats were hand-fed different ratios of legume hay and concentrate (n = 60), or combinations of three browse species and concentrate (n = 91), while for the second (n = 153), data were collected during 10 observation days with grazing goats. On these days, pairs of dietary information and feces were derived from one observed goat but feces were also sampled from 12 to 15 resident goats grazing in the same paddock. Each dataset served for the validation of NIRS equations established with the other. For CP, in dataset 1, the values of R2 for calibration were 0.98 (all data) and 0.90 (only browse) and the standard error of cross validation (SECV, internal validation) was 0.50%. When samples from dataset 2 were used for external validation, R 2 varied between 0.55 (all goats) and 0.85 (observed goats) and the standard error of prediction (SEP, external validation) was 2.2 and 1.6% in the same order. Using dataset 1, we determined that the condition for 95% confidence that two goats ate the same diet was a Mahalanobis distance (H) of less than 0.5 between their fecal spectra. Using dataset 2, the values of R2 for calibration varied between 0.90 (all goats) and 0.92 (only observed goats and residents with H < 0.5 from their fecal spectra) and SECV was approximately 0.50%. When samples from dataset 1 were used for external validation, R2 was 0.47 (calibrations with all goats) and 0.50 (only observed goats and resident goats with H < 0.5 from the observed goats), and SEP varied between 0.8 and 1.2% in the same order. Lower R2 and accuracy were found for in vitro DMD than those for CP. A dataset containing feces and dietary information from observed goats seems to be accurate and cost-effective to implement fecal NIRS determination of dietary CP and DM digestibility in grazing goats. 2005 Elsevier B.V. All rights reserved.
