Using faecal near-infrared spectroscopy (FNIRS) to estimate nutrient digestibility and chemical composition of diets and faeces of growing pigs

Near-infrared spectroscopy has been successfully used in predicting nutrient digestibility and other parameters, e.g. diet composition from the spectra of faeces samples in herbivores (FNIRS). The objective of the present study was to investigate if the chemical composition of diets and faeces, as well as the apparent digestibility of organic matter (OM), crude protein (CP), neutral-detergent and acid-detergent fibre (aNDFom, ADFom) and crude fibre (CF) can be predicted with sufficient accuracy from spectra of pigs' faeces. Faecal samples (n = 202), together with detailed information about the chemical composition of these samples and the diets supplied, were provided from five very different digestibility trials with growing pigs. Faecal spectra were obtained by NIRS and calibrations were developed using the complete dataset. Accuracy of calibration equations was assessed by means of standard error of cross-validation (SECV) and the RPD value (standard deviation of lab results/SECV). The prediction of the chemical composition of the diet was considered successful, as all calibration equations had RPD values well above 2, the value that indicates sufficient accuracy. The chemical composition of faeces was predicted with good to sufficient accuracy. Calibrations developed to estimate the apparent nutrient digestibility from faecal spectra worked well for OM and CP, with a SECV of 0.024 for both and RPD values of 2.1 and 2.4. This very promising result merits further investigation. The prediction of the apparent digestibility of fibre fractions was less successful. We conclude that the FNIRS approach is feasible for use in pig nutrition research for predicting the chemical composition of diet and faeces, as well as to determine the apparent digestibility of OM and CP in situations where controlled digestibility trials are not possible. (C) 2015 Elsevier B.V. All rights reserved.