Estimating Body Composition in Mice by Near-Infrared Spectrophotometry

Publisher Website
Google Scholar
Abstract
Fat, protein and ash content of mice and other animal species are important response variables measured in genetic, nutritional and physiological studies. Procedures for estimating body composition usually include determination of fat by solvent extraction, N by the Kjeldahl method and ash by combustion. Near-infrared spectrophotometry (NIR) is a potentially rapid technique for estimating all three components simultaneously from a single instrument scan. One group of mice was used for calibration by relating the chemical determinations to optical variables of the wavelengths that provided the best fit for each component based on stepwise multiple linear regression. Independent variables were second derivatives of logarithms of reciprocals of reflectance for each wavelength. Coefficients of determination of calibration were .94, .91 and .64 for percentage of fat, protein and ash. Multiple regression procedures based on correlograms and characteristic spectra yielded comparable coefficients of determination. A nested sampling experiment provided estimates of percentage variation in each constituent associated with 1) individual mice, 2) sampling from total ground tissue, 3) packing of sample in the disk for reading and 4) instrument reading error. Respective percentages were 91.2, .3, 3.9 and 4.6 for fat; 94.4, .2, 2.6 and 2.8 for protein and 40.3, .0, .0 and 59.7 for ash. Repeatabilities based on NIR were .91, .94 and .40 for fat, protein and ash percentage, respectively. A second group of mice was used for evaluating the predictability of the calibration equations. Coefficients of determination of prediction were .92, .88 and .68 for fat, protein and ash percentage, respectively. Near-infrared spectrophotometry appears to be a useful technique for rapidly estimating fat and protein percentage in ground samples of mouse tissue and possibly other species, but seems less useful for predicting ash percentage. Copyright © 1984. American Society of Animal Science. Copyright 1984 by American Society of Animal Science.

This publication has 0 references indexed in Scilit: