Predicting fiber content in herbivore fecal samples using a multispecies NIRS model
Mariana Rossa1, Emmanuel Serrano2, João Carvalho1
1Departamento de Biologia e Centro de Estudos do Ambiente e do Mar (CESAM), Universidade de Aveiro, Campus Universitário de Santiago, Aveiro, Portugal.
Plos One
|January 8, 2025
Summary
New multispecies models accurately predict fecal fiber content (neutral detergent fiber, acid detergent fiber, and acid detergent lignin) in herbivores using near-infrared reflectance spectroscopy. This advances nutritional ecology studies across diverse species.
Area of Science:
- Animal Nutrition and Metabolism
- Nutritional Ecology
- Spectroscopy Applications
Background:
- Fiber content in herbivore feces is crucial for understanding digestive health and nutritional ecology.
- Traditional near-infrared reflectance spectroscopy (NIRS) calibrations for fecal fiber are species-specific, requiring extensive sample sets.
- A multispecies calibration for NIRS would enable broader application across diverse herbivore species.
Purpose of the Study:
- To develop and validate multispecies near-infrared reflectance spectroscopy (NIRS) calibration models for predicting fecal fiber components (neutral detergent fiber, acid detergent fiber, and acid detergent lignin).
- To assess the accuracy of these multispecies models for diverse domestic and wild herbivore species.
- To evaluate the impact of including sodium sulfite in the fiber determination protocol.
Main Methods:
- Developed multispecies NIRS calibration models for fecal neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) using 445 samples from six herbivore species.
- Utilized 80% of samples for calibration and 20% for internal validation.
- Performed external validation using 63 samples from five additional herbivore species not included in the calibration set.
Main Results:
- Highly accurate multispecies NIRS models were developed for NDF, ADF, and ADL (calibration R² ≥ 0.93, validation R² ≥ 0.91).
- External validation showed satisfactory accuracy for NDF (R² 0.86–0.94) and ADF (R² 0.80–0.95), and acceptable accuracy for ADL (R² 0.66–0.89) across different species.
- The inclusion of sodium sulfite in the protocol did not significantly improve fecal fiber prediction accuracy.
Conclusions:
- Multispecies NIRS calibrations provide a highly accurate and versatile tool for assessing fecal fiber content across a wide range of herbivore species.
- These models represent a significant advancement for studying the nutritional ecology of herbivores with varied foraging strategies.
- Future improvements can be achieved by expanding the dataset with more species and geographical locations.
