Compilation of Studies
Animal Feed, Forage, and Pet Food Applications Review
The Current and Future Role of Near Infrared Reflectance Spectroscopy in Animal Nutrition: A Review
While traditional approaches to feed characterization do provide insight into feed quality, they are insufficient in meeting a number of challenges in the animal feed manufacturing industry. There is an emphasis on designing diets using feed formulation that offer minimization of feed cost that meet the nutritional requirements of not only different species of animals, but also different breeds and life cycles. Reduction of negative environmental impact is considered a priority as well. There is an increasing need to predict the output of milk and meat components, such as fat, fatty acids, protein, and lactose. Thus, the industry desires an approach to nutrition based on multiple responses of the animal to the nutrients supplied by the diet. Accurate evaluation methods are also needed regarding national and international regulations and safety needs. Traditional analytical methods have proven to be insufficient for meeting the growing needs of animal feed testing. They are tedious, time-consuming, expensive, only capable of measuring a single parameter at a time, and are ill-suited for testing samples on a large scale. NIR spectroscopy offers the advantages of being fast, non-invasive, the ability to measure multiple parameters with one measurement, and requiring little or no sample preparation while using no toxic chemicals or reagents. In this review paper, an overview of research and applications of NIR spectroscopy in relation to animal feed are examined, with a particular emphasis on forage characterization and digestibility parameters.
Forages provide the basis for ruminant production systems. It is estimated that over 70% of the total worldwide agricultural area is grassland and that forages provide over 90% of the feed energy consumed by herbivorous animals. The first successful use of NIR spectroscopy to predict forage digestibility was done by Karl Norris in 1976, with R2 values of 0.78 and 0.90 for the prediction of dry matter (DM) digestibility in vitro and in vivo. Subsequent research expanded on digestibility parameters like metabolizable energy (ME) and chemical nutritional parameters, such as crude protein and neutral detergent fiber (NDF). Forages tested in documented studies in the first twenty years of research include grass, grass silage, hays, cereal straw, and various mixtures. An important study conducted by Barber et al. in 1990 showed that NIR spectroscopy could predict organic matter digestibility (OMD) in vivo more accurately than commonly used laboratory methods. Another study conducted in Belgium a few years later showed an R2 of 0.79 for OMD in vivo using an NIR calibration as opposed to 0.68, 0.64, 0.53, and 0.40 for cellulose digestibility, rumen fluid-pepsin OMD, acid detergent lignin, and pepsin-cellulase OMD determinations.
Most of the early research on NIR spectroscopy and forages used dried samples and it was demonstrated that water has a profound effect on the spectra of silage in the wavelength range from 1850 nm to 2000 nm. Samples dried in the same oven could show different moisture levels with an effect on the NIR spectra and in order to reduce the sensitivity level from moisture differences, algorithms were used to down weigh the contribution from moisture sensitive wavelength regions. Similarly, variations in particle size can account for large amounts of spectral variance. Different data pretreatments such as Standard Normal Variate-Detrending (SNV-D) and scatter correction were shown to improve calibrations. In 1993, a model was created for predicting the OMD in vivo of grass silage that was agreed to be the standard for use throughout the UK by both government regulators and the feed industry. The model showed a R2 of 0.83 and a Standard Error of Prediction (SEP) of 2.40. At the time, various chemometric algorithms were being used for calibration and this study along with others showed the Partial Least Squares (PLS) algorithm to be most effective. The expense of monochromators led to investigation of using fixed filter instruments for OMD that measured nineteen wavelengths. While the results were not as good as using a full range monochromators, models still showed an improvement over laboratory results.
Other research and applications studies focused on fresh silages that are not dried or milled. The distinction is important as the drying of samples takes time, can potentially change cell walls from high temperature, and can lose volatile acids, alcohols, esters, amines, and ammonia. It became clear that drying could have a profound effect on these nutrients as the focus shifted from describing energy and protein values in terms of digestibility to quantifying nutrients available for absorption. Examples of other parameters studied include total nitrogen, dry nitrogen, fermentation acids, in vitro cell wall degradability characteristics, neutral detergent fiber (NDF), and acid detergent fiber (ADF). Work progressed from silages to the evaluation of concentrates like raw materials and compound feeds. Early studies showed the inherent challenges of using NIR spectroscopy for such analysis. A study in Germany in 1990 showed that proximate analysis for twenty-two different raw feed materials such as wheat, barley, maize, tapioca, and soybean meal was feasible. The extent of the effect of raw material variation was made apparent in such studies and the benefits of local calibrations were discovered, such as having separate calibrations for poultry layer and broiler feeds. Other studied parameters included amino acids, minerals, and identification and authenticity using NIR spectroscopy.
Although this paper is dated and cites studies going back over twenty years, it offers a fascinating view of the foundation for the progress that has been made in using NIR spectroscopy for animal feed and forage analysis. Many of the complex nutritional parameters initially studied in this review have been proven to be measurable by NIR spectroscopy. The extent to which NIR spectroscopy has revolutionized animal feed analysis cannot be understated. It has provided a fast and rapid means for testing in all areas of the process, including raw material analysis, on-line process analysis in feed mills, initial and final quality control checks, and a means for testing for feed being authentic and meeting government regulations. More than anything, the estimation of the actual nutrient supply of the animal as opposed to energy and protein analysis gives farmers and manufacturers a means for determining the output of animal products, allowing for maximum profits while minimizing costs. New technologies and techniques will continue to advance the use of NIR spectroscopy in the animal feed industry.
A Review on the Use of Near-Infrared Spectroscopy for Analyzing Feed Protein Materials
There are many different feed protein materials of both plant and animal origin used in animal feed. Plant origin materials include soybean, rapeseed, cottonseed, peanut, and sunflower seed meals. Animal origin materials include fish, meat, and bone meals. It is estimated that the world consumption of feed protein materials amounts to well over 200 million metric tons annually. Strong safety regulations have been established, necessitating the need for efficient and accurate analytical tools. Traditional methods for such analysis usually entail wet chemistry methods and while accurate, there are shortcomings that are insufficient to meet the growing demands and needs of the industry. Most wet chemistry methods are time-consuming and laborious as well as only having the ability to measure one parameter at a time. NIR spectroscopy offers the advantages of being fast and non-invasive while requiring minimal or no sample preparation. It can also analyze multiple parameters from a single measurement as well as measure a large volume of sample. Standard analysis of parameters of interest often uses a small volume of sample that is insufficient to account for variation within batches. In this review, a summary of advances in NIR spectroscopy for monitoring the quality and safety of feed protein materials is examined.
Analysis of the chemical contents of feed protein materials is crucial for feed manufacturers. It enables them to efficiently arrange formulations while meeting quality control standards and safety regulations at minimal cost. Moisture is a proven parameter that can be measured accurately and efficiently by NIR spectroscopy as O-H bands have very high absorptivity in the NIR spectral range. Numerous studies have determined moisture in various feed protein materials with R2 > 0.90 and due to the high correlation between dry matter and moisture, dry matter is also a proven measurable parameter using NIR spectroscopy. Crude protein is an important index in the nutritive value of feed materials. Like moisture, a number of studies examining crude protein have shown R2 > 0.90, including in soybean meal, rapeseed meal, cottonseed meal, peanut meal, and fish meal. Good predictability of crude protein is expected using NIR methods as a number of absorption bands in the NIR region closely relate to protein. These include the N-H stretch first overtone at 1500 nm to 1530 nm, the aromatic C-H stretch first overtone in the region from 1640 nm to 1680 nm, the carbonyl stretch of the primary amide at 2060 nm, and the N-H/C-H/C O/C-N combination bands from 2168 nm to 2180 nm. Fat is the main component in determining feed energy level. Compared to moisture and crude protein, results of studies have shown fat to be more difficult to predict using NIR spectroscopy, especially in plant materials. However, some correlations are still at R2 > 0.90. It is likely that fat analysis in plant materials is more difficult because of lower fat content. Plant materials used for feed are often by-products of oil extraction, creating a low fat content. In contrast, fish meal has a fat content around 10% and one study showed a very high correlation of R2 = 0.96 when determining fat content in fish meal.
Ash is the inorganic residue remaining after the water and organic material are removed by being subjected to high temperatures. It provides a measure of the total amount of minerals in feed materials. While the residue is mainly made up of mineral elements which have no characteristic NIR absorption bands, ash is a proven parameter for NIR analysis. It is generally agreed that NIR spectroscopy can be used to evaluate mineral contents due to the correlation between minerals and the organic components, either through associations with organic molecules or by the formation of salts which affect the hydrogen bonds in samples. Amino acid content is important as well as a lack of certain amino acids, especially methionine and lysine, can limit the nutritional efficiency of feed materials. Studies have shown that a number of amino acids can accurately be determined from calibrations using NIR spectra and reference methods. However, some amino acids show much better predictive performance than others. Two examples of amino acids which have shown inferior predictions are methinonine and cystine. This is likely due to their easily oxidative properties and inaccurate reference analyses. Research studies proved that the reproducibility error of reference analysis for other amino acids is typically below 2% for most amino acids while methionine is above 3% and cystine is between 5% and 6%. It can be concluded that this reference error is one cause of lower correlation for these two amino acids.
Traceability and authenticity determination of feed materials is important for quality determination as well as public health and safety. Mad cow disease is of particular concern as the most likely route of infection is from feeds that contain low protein levels. While a number of methods exist for analyzing feed ingredients for authenticity and safety, including microscopy, DNA detection, immunochemistry, and mass spectroscopy, these methods are expensive and ill-suited for large scale analysis. NIR spectroscopy has been analyzed in a number of studies and shown to be effective for traceability and authenticity determination of feed materials. Most research is focused on two types of analysis: the detection and quantification of prohibited feed ingredients and detection and quantification of low-cost ingredients as adulterants in fish meal. One example of a prohibited ingredient in compound feed is meat and bone meal (MBM). A number of studies for determining MBM in compound feed and fish meal showed R2 ranging from 0.94 to 0.98. Fish meal is the most nutritive protein feed and therefore the most expansive, making adulteration with cheaper plant based materials a big issue in the industry. Likewise, high correlation with R2 has been shown in numerous studies that adulterated fish meal with soybean meal and similar plant protein feeds. Studies have also shown that adulteration with both MBM and plant protein feeds can be quantified if the concentration of the adulterant is greater than 1%. As new challenges emerge in protein feed analysis, NIR spectroscopy will be further studied and examined to determine its use as an effective analytical tool in the industry.
An Overview on the Use of Near Infrared Spectroscopy (NIRS) on Farms for the Management of Dairy Cows
Dairy farming has been greatly affected by the digital revolution. New technologies are strongly influencing farm management in various ways, such as reducing manual labor, costs, waste output and subsequently, increasing efficiency and profits. “Precision agriculture” is defined as a system run through the collection, integration, and automatic analysis of data from the environment, sensors, and any other third source. When it comes to the livestock sector, the similar term “Precision Livestock Farming” (PLF) has developed the whole sector from cattle to pigs to poultry and most especially dairy cattle, which has proved to be the sector that is most affected by variation in nutritional intake in terms of animal welfare, environmental impact, and profitability. Feeding represents well over half the milk production costs for dairy cattle farms. Therefore, it is very important to avoid mistakes both in the formulation and in the distribution of the ration, since they may involve digestive process inefficiency, production losses, worsening production quality, reduction of reproductive performance, increase of production costs, greater waste and environmental impact, animal health problems, worsening of well-being, and greater drug consumption. While the feasibility of using NIR spectroscopy for measuring nutritional parameters in animal feed has long been proven, technological advances have facilitated its use as both an on-site and field measurement in the farm as well as a tool in animal feed manufacturing. In this review, an examination of the use of NIR spectroscopy at the barn level is examined with a particular focus on the use of portable instruments.
There are five particular steps in the milk manufacturing process where portable NIR instruments are used. They are as follows: analysis of the chemical composition of raw materials in the field and the loading phase of the mixer wagon, analysis of the total mixed ration (TMR) for both chemical and physical properties, analysis of TME evaluation indices, analysis of the chemical composition of feces, slurry, and manure, and on-line analysis of milk quality in the milking parlor. Numerous studies have been conducted examining chemical composition of fodder and silages used to feed dairy cattle. Initial research was conducted on dried samples and with advances in chemometric analysis, data processing, and improved instrument hardware, a number of applications have now been developed to successfully analyze fresh products. Proven parameters of interest include dry matter, total nitrogen, insoluble nitrogen, ammonia nitrogen, pH, lactic acid, and acetic acid. Fermentation parameters in particular proved difficult to measure in dried samples but show better results in fresh samples. With regard to TMR analysis, diet preparation allows animals to ingest balanced meals in terms of nutritional composition and physical structure. Animals who are fed poor TMR can show adverse health effects, reduced milk production, and lower milk quality. Studies using portable NIR instruments for such analysis have shown good results for a number of parameters, including standard proximate analysis, lipids, starch, ADF, and NDF. Accurate DM monitoring is crucial as any variation in DM leads to a risk of overfeeding or underfeeding the animals.
After proper formation of TMR, it is very important to evaluate homogeneity and selection index. For a mixture to be defined as truly homogenous, it must contain the individual ingredients by both chemical composition and physical structure with uniform distribution and without any undesirable or harmful deficiencies or excesses. Reasons for non-homogeneity in TMR distribution may be attributable to various factors, including type of mixer wagon and integrity of the cutting elements, the loading sequence, the mixing time, attention of the operator, and possible modifications of raw materials. Although this type of analysis using NIR spectroscopy is relatively new, one study developed an algorithm capable of providing estimated homogeneity for TMR in both milk and beef cattle. A number of factors were examined including the effects of homogeneity in TMR prepared with the addition of water as opposed to dry TMR. Although more research would be needed before real implementation of this analysis, the results were promising and could potentially be a powerful tool in increasing precision feeding benefits.
NIR spectroscopy has been used to analyze the chemical composition of slurry and manure. Chemical composition of feces is useful for determining the quantity and quality of nutrients that are useful for fertilization, the quality of the material for composting processes in terms of implementing efficient emission reduction strategies, and information about the digestibility of the ration. Examples of parameters analyzed in manure using NIR spectroscopy include pH, total N, total C, organic C, C:N ratio, P, available P, S, K, Na, and nitrate + nitrite. Correlation for these parameters is generally good with the exceptions of nitrate + nitrite, available P, and Na. One particular study showed correlation of R2 > 0.94 for moisture, total C, and total N. Digestibility parameters are proven as well from fecal analysis using NIR spectroscopy. Numerous studies have shown the feasibility of measuring both chemical and physical parameters of interest in dairy products as well, including milk.
It must be noted that studies have also been conducted comparing analysis of using portable and traditional laboratory instruments for NIR analysis. Portable instruments generally show worse calibration results than benchtop instruments due to a number of factors. These include limited wavelength range, lower resolution, and lower signal-to-noise ratio. However, the potential benefits of the portability of NIR instruments cannot be understated. Advancements and research to improve both the efficiency and cost of these instruments is on-going and as continued new challenges occur in the farm and dairy cattle industries, the use of portable NIR instruments will play an important role as an analytical tool designed to meet these challenges.
Direct and Indirect Means of Predicting Forage Quality Through Near Infrared Reflectance Spectroscopy
Animal nutritionists have long recognized the importance of determining the nutritive value in feeds and forages fed to livestock and have used analytical chemistry to determine protein, energy, and mineral content. However, these methods can be expensive, time-consuming, and ill-suited for large scale measurement of parameters of interest. NIR spectroscopy is a proven alternative method for analysis of feed and forage materials fed to livestock. While the characterization of feeds and forages fed to an animal is a relatively simple task, it is more difficult to quantify the nutritional value of a diet obtained by a grazing animal. In this review, both the direct analysis of forages and indirect analysis of feces are examined for major chemical and physical parameters of interest in forage are examined.
Ash is defined as the inorganic portion of a compound after the organic compounds are burned off and generally comprises the total mineral content. It has successfully been determined in a number of forage types using NIR spectroscopy. Protein, nitrogen, and related compounds are also proven constituents that can be measured, especially in the case of crude protein which has a relatively high concentration in feed materials. While fiber parameters like ADF and NDF are considered “properties” of forages and not constituents, they can still be estimated using NIR spectroscopy due to variation in the C-H and O-H bonds. Lignin can be estimated as well but generally has higher calibration error than protein and ADF. Lipids and ether extract do show C-H absorptions around 1725 nm and 2310 nm but calibration models are generally poor, likely due to low concentration and small variance in forage. Some micro minerals have been successfully measured using NIR spectroscopy, such as Ca, P, Mg, and K. While minerals do not have any direct NIR absorption bands, they can affect organic molecules that absorb in the NIR region so an indirect correlation for certain minerals is possible. However, models using an indirect correlation of this nature must be carefully examined and validated. Digestibility parameters are proven as well and some anti-quality components have been successfully measured using NIR spectroscopy, such as total alkaloid concentration in poisonous plants and condensed tannins.
Another method for determining diet quality by using NIR spectroscopy is fecal analysis. As the diet chemistry of free-ranging animals changes, the by-products of ingestion also change. These include plant residue, microbial bodies, secondary metabolites, and slough tissue. The behavior of these secondary products in the feces may be related to characteristics of the primary product. Studies examining the potential of fecal analysis through NIR spectroscopy pair fecal spectra with the known chemical and physical parameters of the dietary nutritional quality and subsequent validation of the calibration equations with unknown samples. While natural variation in NIR spectra of agricultural products requires a broad inclusion of different sample types, the need for this is even more pronounced for fecal analysis as highly diverse temporal, spatial, species, environmental, and landscape conditions can occur in the forage consumed by animals. Good results have been obtained in measuring crude protein and digestibility parameters from feces, with some studies showing R2 as high as 0.99 for crude protein and 0.95 for different digestibility parameters. Species used in these studies include cattle, goats, deer, and elk. While further research is needed and the challenges of the calibration work are extensive for fecal analysis, the potential of using NIR spectroscopy to indirectly monitor dietary intake of ruminants from feces has been demonstrated in numerous successful studies.