Large-scale phenotyping in dairy sector using milk MIR spectra: Key factors affecting the quality of predictions.

Methods and technologies enabling the estimation at large scale of important traits for the dairy sector are of great interest. Those phenotypes are necessary to improve herd management, animal genetic evaluation, and milk quality control. In the recent years, the research was very active to predict new phenotypes from the mid-infrared (MIR) analysis of milk. Models were developed to predict phenotypes such as fine milk composition, milk technological properties or traits related to cow health, fertility and environmental impact. Most of models were developed within research contexts and often not designed for routine use. The implementation of models at a large scale to predict new traits of interest brings new challenges as the factors influencing the robustness of models are poorly documented. The first objective of this work is to highlight the impact on prediction accuracy of factors such as the variability of the spectral and reference data, the spectral regions used and the complexity of models. The second objective is to emphasize methods and indicators to evaluate the quality of models and the quality of predictions generated under routine conditions. The last objective is to outline the issues and the solutions linked with the use and transfer of models on large number of instruments. Based on partial least square regression and 10 datasets including milk MIR spectra and reference quantitative values for 57 traits of interest, the impact of the different factors is illustrated by evaluating the influence on the validation root mean square error of prediction (RMSEP). In the displayed examples, all factors, when well set up, increase the quality of predictions, with an improvement of the RMSEP ranging from 12% to 43%. This work also aims to underline the need for and the complementarity between different validation procedures, statistical parameters and quality assurance methods. Finally, when using and transferring models, the impact of the spectral standardization on the prediction reproducibility is highlighted with an improvement up to 86% with the tested models, and the monitoring of individual spectrometer stability over time appears essential. This list inspired from our experience is of course not exhaustive. The displayed results are only examples and not general rules and other aspects play a role in the quality of final predictions. However, this work highlights good practices, methods and indicators to increase and evaluate quality of phenotypes predicted at a large scale. The results obtained argue for the development of guidelines at international levels, as well as international collaborations in order to constitute large and robust datasets and enable the use of models in routine conditions.

Challenges for cocaine detection in smuggling samples.

Screening of seized cocaine powders is routinely performed by means of colour tests. An alternative fast screening technique is Mid-InfraRed (MIR) spectroscopy. In the context of smuggling cases, however, drugs are often processed to circumvent detection. In this study, the current screening techniques (cocaine colour test and MIR spectroscopy using libraries and chemometrics) were applied to five smuggling cases. For each case, all samples were first screened with a cocaine colour test and MIR analysis, followed by confirmation analyses with GC-MS and GC-FID to identify and quantify cocaine and cutting agents. Finally, Scanning Electron Microscopy-Energy Dispersive X-ray spectroscopy (SEM-EDX) analyses were performed for additional characterization. All smuggling samples tested negative, both on-site as in the laboratory, for cocaine with the cocaine colour test. Four smuggling cases consisted of coloured samples. Consequently the colour test result was influenced because discolouration of the test showed almost the same colour as the colour of the powders (brown, green, red or black). In contrast, the (coloured) powders could be measured with MIR, but the MIR spectra showed no hit for cocaine using a reference library search. Moreover, cocaine was not detected in four out of the five cases after application of a chemometric classification model. GC-MS analysis, the golden standard for identification, resulted in a positive identification of cocaine in all cases. These samples contained cocaine ranging between 0.8w% and 35w%. Taking into account the results of the screening, the chromatographic and the SEM-EDX analyses, it could be presumed that cocaine was masked. False negative screening results were caused by chemically modified cocaine and/or cocaine mixed with coloured powders. In additional experiments, a sample extraction step prior to the screening techniques was performed. Two sample preparation methods (acetone and ethyl acetate) were tested and resulted in a positive screening of cocaine with the colour test and/or MIR spectroscopy. It can be concluded that the outcome of screening techniques such as colour tests and MIR spectroscopy is only presumptive and should always be confirmed.

Evaluation of a calibration transfer between a bench top and portable Mid-InfraRed spectrometer for cocaine classification and quantification.

A portable Fourier Transform Mid-InfraRed (FT-MIR) spectrometer using Attenuated Total Reflectance (ATR) sampling is used for daily routine screening of seized powders. Earlier, ATR-FT-MIR combined with Support Vector Machines (SVM) algorithms resulted in a significant improvement of the screening method to a reliable and straightforward classification and quantification tool for both cocaine and levamisole. However, can this tool be transferred to new (hand-held) devices, without loss of the extensive data set? The objective of this study was to perform a calibration transfer between a newly purchased bench top (BT) spectrometer and a portable (P) spectrometer with existing calibration models. Both instruments are from the same brand and have identical characteristics and acquisition parameters (FT instrument, resolution of 4 cm-1 and wavenumber range 4000 to 500 cm-1). The original SVM classification model (n = 515) and SVM quantification model (n = 378) were considered for the transfer trial. Three calibration transfer strategies were assessed: 1) adjustment of slope and bias; 2) correction of spectra from the new instrument BT to P using Piecewise Direct Standardization (PDS) and 3) building a new mixed instrument model with spectra of both instruments. For each approach, additional cocaine powders were measured (n = 682) and the results were compared with GC-MS and GC-FID. The development of a mixed instrument model was the most successful in terms of performance. The future strategy of a mixed model allows applying the models, developed in the laboratory, to portable instruments that are used on-site, and vice versa. The approach offers opportunities to exchange data within a network of forensic laboratories using other FT-MIR spectrometers.

Practical tool for sampling and fast analysis of large cocaine seizures.

Large quantities of illicit drugs are frequently seized by law enforcement. In such cases, a representative number of samples needs to be quickly examined prior to destruction. No procedure has yet been set up which rapidly provides information regarding the homogeneity of the samples, the presence of controlled substances, and the degree of purity. This study establishes a protocol for fast analysis of cocaine and its most common cutting agent, levamisole, in large seizures. The protocol is based on a hypergeometric sampling approach combined with Fourier-transform infrared (FTIR) spectrometry and support vector machines (SVM) algorithms as analysis methods. To demonstrate the practical use of this approach, 5 large cocaine seizures (consisting between 45 and 85 units) were analysed simultaneously with gas chromatography-mass spectrometry (GC-MS), gas chromatography-flame ionisation detector (GC-FID), and a portable FTIR spectrometer using attenuated total reflectance (ATR) sampling combined with SVM models. According to the hypergeometric sampling plan of the guidelines of the Drugs Working Group (DWG) of the European Network of Forensic Science Institutes (ENFSI), the required number of subsamples ranged between 19 and 23. Considering the identification analyses, the SVM models detected cocaine and levamisole in all subsamples of Cases 1 to 5 (100% correct classification), which was confirmed by GC-MS analysis. Considering the quantification analyses, the SVM models were able to estimate the cocaine and levamisole content in each subsample, compared to GC-FID data. The developed strategy is easy, cost effective, and provides immediate information about both the presence and concentration of cocaine and levamisole. By using this new strategy, the number of confirmation analyses with laborious and expensive chromatographic techniques could be significantly reduced.

Standardization of milk mid-infrared spectrometers for the transfer and use of multiple models.

An increasing number of models are being developed to provide information from milk Fourier transform mid-infrared (FT-MIR) spectra on fine milk composition, technological properties of milk, or even cows' physiological status. In this context, and to take advantage of these existing models, the purpose of this work was to evaluate whether a spectral standardization method can enable the use of multiple equations within a network of different FT-MIR spectrometers. The piecewise direct standardization method was used, matching "slave" instruments to a common reference, the "master." The effect of standardization on network reproducibility was assessed on 66 instruments from 3 different brands by comparing the spectral variability of the slaves and the master with and without standardization. With standardization, the global Mahalanobis distance from the slave spectra to the master spectra was reduced on average from 2,655.9 to 14.3, representing a significant reduction of noninformative spectral variability. The transfer of models from instrument to instrument was tested using 3 FT-MIR models predicting (1) the quantity of daily methane emitted by dairy cows, (2) the concentration of polyunsaturated fatty acids in milk, and (3) the fresh cheese yield. The differences, in terms of root mean squared error, between master predictions and slave predictions were reduced after standardization on average from 103 to 17 g/d, from 0.0315 to 0.0045 g/100 mL of milk, and from 2.55 to 0.49 g of curd/100 g of milk, respectively. For all the models, standard deviations of predictions among all the instruments were also reduced by 5.11 times for methane, 5.01 times for polyunsaturated fatty acids, and 7.05 times for fresh cheese yield, showing an improvement of prediction reproducibility within the network. Regarding the results obtained, spectral standardization allows the transfer and use of multiple models on all instruments as well as the improvement of spectral and prediction reproducibility within the network. The method makes the models universal, thereby offering opportunities for data exchange and the creation and use of common robust models at an international level to provide more information to the dairy sector from direct analysis of milk.

Regression models based on new local strategies for near infrared spectroscopic data.

In this work, a comparative study of two novel algorithms to perform sample selection in local regression based on Partial Least Squares Regression (PLS) is presented. These methodologies were applied for Near Infrared Spectroscopy (NIRS) quantification of five major constituents in corn seeds and are compared and contrasted with global PLS calibrations. Validation results show a significant improvement in the prediction quality when local models implemented by the proposed algorithms are applied to large data bases.

Development of Fourier transform mid-infrared calibrations to predict acetone, ß-hydroxybutyrate, and citrate contents in bovine milk through a European dairy network.

To manage negative energy balance and ketosis in dairy farms, rapid and cost-effective detection is needed. Among the milk biomarkers that could be useful for this purpose, acetone and ß-hydroxybutyrate (BHB) have been proved as molecules of interest regarding ketosis and citrate was recently identified as an early indicator of negative energy balance. Because Fourier transform mid-infrared spectrometry can provide rapid and cost-effective predictions of milk composition, the objective of this study was to evaluate the ability of this technology to predict these biomarkers in milk. Milk samples were collected in commercial and experimental farms in Luxembourg, France, and Germany. Acetone, BHB, and citrate contents were determined by flow injection analysis. Milk mid-infrared spectra were recorded and standardized for all samples. After edits, a total of 548 samples were used in the calibration and validation data sets for acetone, 558 for BHB, and 506 for citrate. Acetone content ranged from 0.020 to 3.355mmol/L with an average of 0.103mmol/L; BHB content ranged from 0.045 to 1.596mmol/L with an average of 0.215mmol/L; and citrate content ranged from 3.88 to 16.12mmol/L with an average of 9.04mmol/L. Acetone and BHB contents were log-transformed and a part of the samples with low values was randomly excluded to approach a normal distribution. The 3 edited data sets were then randomly divided into a calibration data set (3/4 of the samples) and a validation data set (1/4 of the samples). Prediction equations were developed using partial least square regression. The coefficient of determination (R(2)) of cross-validation was 0.73 for acetone, 0.71 for BHB, and 0.90 for citrate with root mean square error of 0.248, 0.109, and 0.70mmol/L, respectively. Finally, the external validation was performed and R(2) obtained were 0.67 for acetone, 0.63 for BHB, and 0.86 for citrate, with respective root mean square error of validation of 0.196, 0.083, and 0.76mmol/L. Although the practical usefulness of the equations developed should be further verified with other field data, results from this study demonstrated the potential of Fourier transform mid-infrared spectrometry to predict citrate content with good accuracy and to supply indicative contents of BHB and acetone in milk, thereby providing rapid and cost-effective tools to manage ketosis and negative energy balance in dairy farms.

Capitalizing on fine milk composition for breeding and management of dairy cows.

The challenge of managing and breeding dairy cows is permanently adapting to changing production circumstances under socio-economic constraints. If managing and breeding address different timeframes of action, both need relevant phenotypes that allow for precise monitoring of the status of the cows, and their health, behavior, and well-being as well as their environmental impact and the quality of their products (i.e., milk and subsequently dairy products). Milk composition has been identified as an important source of information because it could reflect, at least partially, all these elements. Major conventional milk components such as fat, protein, urea, and lactose contents are routinely predicted by mid-infrared (MIR) spectrometry and have been widely used for these purposes. But, milk composition is much more complex and other nonconventional milk components, potentially predicted by MIR, might be informative. Such new milk-based phenotypes should be considered given that they are cheap, rapidly obtained, usable on a large scale, robust, and reliable. In a first approach, new phenotypes can be predicted from MIR spectra using techniques based on classical prediction equations. This method was used successfully for many novel traits (e.g., fatty acids, lactoferrin, minerals, milk technological properties, citrate) that can be then useful for management and breeding purposes. An innovation was to consider the longitudinal nature of the relationship between the trait of interest and the MIR spectra (e.g., to predict methane from MIR). By avoiding intermediate steps, prediction errors can be minimized when traits of interest (e.g., methane, energy balance, ketosis) are predicted directly from MIR spectra. In a second approach, research is ongoing to detect and exploit patterns in an innovative manner, by comparing observed with expected MIR spectra directly (e.g., pregnancy). All of these traits can then be used to define best practices, adjust feeding and health management, improve animal welfare, improve milk quality, and mitigate environmental impact. Under the condition that MIR data are available on a large scale, phenotypes for these traits will allow genetic and genomic evaluations. Introduction of novel traits into the breeding objectives will need additional research to clarify socio-economic weights and genetic correlations with other traits of interest.

Origin identification of dried distillers grains with solubles using attenuated total reflection Fourier transform mid-infrared spectroscopy after in situ oil extraction.

The ban on using processed animal proteins in feedstuffs led the feed sector to look for other sources of protein. Dried distillers grains with solubles (DDGS) could be considered as an important source in this regard. They are imported into Europe mainly for livestock feed. Identifying their origin is essential when labelling is missing and for feed safety, particularly in a crisis situation resulting from contamination. This study investigated applying attenuated total reflection Fourier transform mid-infrared spectroscopy (ATR-FT-MIR) to the oil fraction extracted from samples in situ in order to identify the origin of DDGS. The use of spectroscopic and chemometric tools enabled the botanical and geographical origins of DDGS, as well as the industrial process used to produce them, to be identified. The models developed during the study provided a classification higher than 95% using an external validation set.

Hot topic: Innovative lactation-stage-dependent prediction of methane emissions from milk mid-infrared spectra.

The main goal of this study was to develop, apply, and validate a new method to predict an indicator for CH4 eructed by dairy cows using milk mid-infrared (MIR) spectra. A novel feature of this model was the consideration of lactation stage to reflect changes in the metabolic status of the cow. A total of 446 daily CH4 measurements were obtained using the SF6 method on 142 Jersey, Holstein, and Holstein-Jersey cows. The corresponding milk samples were collected during these CH4 measurements and were analyzed using MIR spectroscopy. A first derivative was applied to the milk MIR spectra. To validate the novel calibration equation incorporating days in milk (DIM), 2 calibration processes were developed: the first was based only on CH4 measurements and milk MIR spectra (independent of lactation stage; ILS); the second included milk MIR spectra and DIM information (dependent on lactation stage; DLS) by using linear and quadratic modified Legendre polynomials. The coefficients of determination of ILS and DLS equations were 0.77 and 0.75, respectively, with standard error of calibration of 63g/d of CH4 for both calibration equations. These equations were applied to 1,674,763 milk MIR spectra from Holstein cows in the first 3 parities and between 5 and 365 DIM. The average CH4 indicators were 428, 444, and 448g/d by ILS and 444, 467, and 471g/d by DLS for cows in first, second, and third lactation, respectively. Behavior of the DLS indicator throughout the lactations was in agreement with the literature with values increasing between 0 and 100 DIM and decreasing thereafter. Conversely, the ILS indicator of CH4 emission decreased at the beginning of the lactation and increased until the end of the lactation, which differs from the literature. Therefore, the DLS indicator seems to better reflect biological processes that drive CH4 emissions than the ILS indicator. The ILS and DLS equations were applied to an independent data set, which included 59 respiration chamber measurements of CH4 obtained from animals of a different breed across a different production system. Results indicated that the DLS equation was much more robust than the ILS equation allowing development of indicators of CH4 emissions by dairy cows. Integration of DIM information into the prediction equation was found to be a good strategy to obtain biologically meaningful CH4 values from lactating cows by accounting for biological changes that occur throughout the lactation.

Standardization of milk mid-infrared spectra from a European dairy network.

The goal of this study was to find a procedure to standardize dairy milk mid-infrared spectra from different Fourier transform mid-infrared spectrophotometers (different brands or models) inside a European dairy network to create new farm-management indicators (e.g., fertility, health, feed, environmental impact) based on milk infrared spectra. This step is necessary to create common spectral databases, allowing the building of statistical tools, to be used by all instruments of the network. The method used was piecewise direct standardization (PDS), which matches slave-instrument spectra on master-instrument spectra. To evaluate the possibility of using common equations on different instruments, the PDS method was tested on a set of milk samples measured on each machine, and an equation predicting fat content of milk is applied on all. Regressions were performed between master and slaves fat predictions, before and after PDS. Bias and root mean square error between predictions were decreased after PDS, respectively, from 0.3781 to 0.0000 and from 0.4609 to 0.0156 (g of fat/100mL of milk). The stability over time of these results was confirmed by an application of the coefficients created by PDS 1 mo later on the slave spectra. These preliminary results showed that the PDS method permits a reduction of the inherent spectral variability between instruments, allowing the merging of Fourier transform mid-infrared milk spectra from different instruments into a common database, the creation of new types of dairy farm management indicators, and the use of these common calibrations for all Fourier transform mid-infrared instruments of the European dairy network.

Validation and transferability study of a method based on near-infrared hyperspectral imaging for the detection and quantification of ergot bodies in cereals.

In recent years, near-infrared (NIR) hyperspectral imaging has proved its suitability for quality and safety control in the cereal sector by allowing spectroscopic images to be collected at single-kernel level, which is of great interest to cereal control laboratories. Contaminants in cereals include, inter alia, impurities such as straw, grains from other crops, and insects, as well as undesirable substances such as ergot (sclerotium of Claviceps purpurea). For the cereal sector, the presence of ergot creates a high toxicity risk for animals and humans because of its alkaloid content. A study was undertaken, in which a complete procedure for detecting ergot bodies in cereals was developed, based on their NIR spectral characteristics. These were used to build relevant decision rules based on chemometric tools and on the morphological information obtained from the NIR images. The study sought to transfer this procedure from a pilot online NIR hyperspectral imaging system at laboratory level to a NIR hyperspectral imaging system at industrial level and to validate the latter. All the analyses performed showed that the results obtained using both NIR hyperspectral imaging cameras were quite stable and repeatable. In addition, a correlation higher than 0.94 was obtained between the predicted values obtained by NIR hyperspectral imaging and those supplied by the stereo-microscopic method which is the reference method. The validation of the transferred protocol on blind samples showed that the method could identify and quantify ergot contamination, demonstrating the transferability of the method. These results were obtained on samples with an ergot concentration of 0.02% which is less than the EC limit for cereals (intervention grains) destined for humans fixed at 0.05%.

Potential use of milk mid-infrared spectra to predict individual methane emission of dairy cows.

This study investigates the feasibility to predict individual methane (CH(4)) emissions from dairy cows using milk mid-infrared (MIR) spectra. To have a large variability of milk composition, two experiments were conducted on 11 lactating Holstein cows (two primiparous and nine multiparous). The first experiment aimed to induce a large variation in CH(4) emission by feeding two different diets: the first one was mainly composed of fresh grass and sugar beet pulp and the second one of maize silage and hay. The second experiment consisted of grass and corn silage with cracked corn, soybean meal and dried pulp. For each milking period, the milk yields were recorded twice daily and a milk sample of 50 ml was collected from each cow and analyzed by MIR spectrometry. Individual CH(4) emissions were measured daily using the sulfur hexafluoride method during a 7-day period. CH(4) daily emissions ranged from 10.2 to 47.1 g CH(4)/kg of milk. The spectral data were transformed to represent an average daily milk spectrum (AMS), which was related to the recorded daily CH(4) data. By assuming a delay before the production of fermentation products in the rumen and their use to produce milk components, five different calculations were used: AMS at days 0, 0.5, 1, 1.5 and 2 compared with the CH(4) measurement. The equations were built using Partial Least Squares regression. From the calculated R(2)(cv), it appears that the accuracy of CH(4) prediction by MIR changed in function of the milking days. In our experimental conditions, the AMS at day 1.5 compared with the measure of CH(4) emissions gave the best results. The R(2) and s.e. of the cross-validation were equal to 0.79 and 5.14 g of CH(4)/kg of milk. The multiple correlation analysis performed in this study showed the existence of a close relationship between milk fatty acid (FA) profile and CH(4) emission at day 1.5. The lower R(2) (R(2) = 0.76) obtained between FA profile and CH(4) emission compared with the one corresponding to the obtained calibration (R(2)(c) = 0.87) shows the interest to apply directly the developed CH(4) equation instead of the use of correlations between FA and CH(4). In conclusion, our preliminary results suggest the feasibility of direct CH(4) prediction from milk MIR spectra. Additional research has the potential to improve the calibrations even further. This alternative method could be useful to predict the individual CH(4) emissions at farm level or at the regional scale and it also could be used to identify low-CH(4)-emitting cows.

Mid-infrared prediction of lactoferrin content in bovine milk: potential indicator of mastitis.

Lactoferrin (LTF) is a milk glycoprotein favorably associated with the immune system of dairy cows. Somatic cell count is often used as an indicator of mastitis in dairy cows, but knowledge on the milk LTF content could aid in mastitis detection. An inexpensive, rapid and robust method to predict milk LTF is required. The aim of this study was to develop an equation to quantify the LTF content in bovine milk using mid-infrared (MIR) spectrometry. LTF was quantified by enzyme-linked immunosorbent assay (ELISA), and all milk samples were analyzed by MIR. After discarding samples with a coefficient of variation between 2 ELISA measurements of more than 5% and the spectral outliers, the calibration set consisted of 2499 samples from Belgium (n = 110), Ireland (n = 1658) and Scotland (n = 731). Six statistical methods were evaluated to develop the LTF equation. The best method yielded a cross-validation coefficient of determination for LTF of 0.71 and a cross-validation standard error of 50.55 mg/l of milk. An external validation was undertaken using an additional dataset containing 274 Walloon samples. The validation coefficient of determination was 0.60. To assess the usefulness of the MIR predicted LTF, four logistic regressions using somatic cell score (SCS) and MIR LTF were developed to predict the presence of mastitis. The dataset used to build the logistic regressions consisted of 275 mastitis records and 13 507 MIR data collected in 18 Walloon herds. The LTF and the interaction SCS × LTF effects were significant (P < 0.001 and P = 0.02, respectively). When only the predicted LTF was included in the model, the prediction of the presence of mastitis was not accurate despite a moderate correlation between SCS and LTF (r = 0.54). The specificity and the sensitivity of models were assessed using Walloon data (i.e. internal validation) and data collected from a research herd at the University of Wisconsin - Madison (i.e. 5886 Wisconsin MIR records related to 93 mastistis events - external validation). Model specificity was better when LTF was included in the regression along with SCS when compared with SCS alone. Correct classification of non-mastitis records was 95.44% and 92.05% from Wisconsin and Walloon data, respectively. The same conclusion was formulated from the Hosmer and Lemeshow test. In conclusion, this study confirms the possibility to quantify an LTF indicator from milk MIR spectra. It suggests the usefulness of this indicator associated to SCS to detect the presence of mastitis. Moreover, the knowledge of milk LTF could also improve the milk nutritional quality.

Online detection and quantification of ergot bodies in cereals using near infrared hyperspectral imaging.

The occurrence of ergot bodies (sclerotia of Claviceps purpurea) in cereals presents a high toxicity risk for animals and humans due to the alkaloid content. To reduce this risk, the European Commission fixed an ergot concentration limit of 0.1% in all feedstuffs containing unground cereals, and a limit of 0.05% in 'intervention' cereals destined for humans. This study sought to develop a procedure based on near infrared hyperspectral imaging and multivariate image analysis to detect and quantify ergot contamination in cereals. Hyperspectral images were collected using an NIR hyperspectral line scan combined with a conveyor belt. All images consisted of lines of 320 pixels that were acquired at 209 wavelength channels (1100-2400 nm). To test the procedure, several wheat samples with different levels of ergot contamination were prepared. The results showed a correlation higher than 0.99 between the predicted values obtained using chemometric tools such as partial least squares discriminant analysis or support vector machine and the reference values. For a wheat sample with a level of ergot contamination as low as 0.01 %, it was possible to identify groups of pixels detected as ergot to conclude that the sample was contaminated. In addition, no false positives were obtained with non-contaminated samples. The limit of detection was found to be 145 mg/kg and the limit of quantification 341 mg/kg. The reproducibility tests of the measurements performed over several weeks showed that the results were always within the limits allowed. Additional studies were done to optimise the parameters in terms of number of samples analysed per unit of time or conveyor belt speed. It was shown that ergot can be detected using a speed of 1-100 mm/s and that a sample of 250 g can be analysed in 1 min.

Comparison of various chemometric approaches for large near infrared spectroscopic data of feed and feed products.

In the present study, different multivariate regression techniques have been applied to two large near-infrared data sets of feed and feed ingredients in order to fulfil the regulations and laws that exist about the chemical composition of these products. The aim of this paper was to compare the performances of different linear and nonlinear multivariate calibration techniques: PLS, ANN and LS-SVM. The results obtained show that ANN and LS-SVM are very powerful methods for non-linearity but LS-SVM can also perform quite well in the case of linear models. Using LS-SVM an improvement of the RMS for independent test sets of 10% is obtained in average compared to ANN and of 24% compared to PLS.

Mid-infrared prediction of bovine milk fatty acids across multiple breeds, production systems, and countries.

Increasing consumer concern exists over the relationship between food composition and human health. Because of the known effects of fatty acids on human health, the development of a quick, inexpensive, and accurate method to directly quantify the fatty acid (FA) composition in milk would be valuable for milk processors to develop a payment system for milk pertinent to their customer requirements and for farmers to adapt their feeding systems and breeding strategies accordingly. The aim of this study was (1) to confirm the ability of mid-infrared spectrometry (MIR) to quantify individual FA content in milk by using an innovative procedure of sampling (i.e., samples were collected from cows belonging to different breeds, different countries, and in different production systems); (2) to compare 6 mathematical methods to develop robust calibration equations for predicting the contents of individual FA in milk; and (3) to test interest in using the FA equations developed in milk as basis to predict FA content in fat without corrections for the slope and the bias of the developed equations. In total, 517 samples selected based on their spectral variability in 3 countries (Belgium, Ireland, and United Kingdom) from various breeds, cows, and production systems were analyzed by gas chromatography (GC). The samples presenting the largest spectral variability were used to calibrate the prediction of FA by MIR. The remaining samples were used to externally validate the 28 FA equations developed. The 6 methods were (1) partial least squares regression (PLS); (2) PLS+repeatability file (REP); (3) first derivative of spectral data+PLS; (4) first derivative+REP+PLS; (5) second derivative of spectral data+PLS; and (6) second derivative+REP+PLS. Methods were compared on the basis of the cross-validation coefficient of determination (R2cv), the ratio of standard deviation of GC values to the standard error of cross-validation (RPD), and the validation coefficient of determination (R2v). The third and fourth methods had, on average, the highest R2cv, RPD, and R2v. The final equations were built using all GC and the best accuracy was observed for the infrared predictions of C4:0, C6:0, C8:0, C10:0, C12:0, C14:0, C16:0, C18:0, C18:1 trans, C18:1 cis-9, C18:1 cis, and for some groups of FA studied in milk (saturated, monounsaturated, unsaturated, short-chain, medium-chain, and long-chain FA). These equations showed R2cv greater than 0.95. With R2cv equal to 0.85, the MIR prediction of polyunsaturated FA could be used to screen the cow population. As previously published, infrared predictions of FA in fat are less accurate than those developed from FA content in milk (g/dL of milk) and no better results were obtained by using milk FA predictions if no corrections for bias and slope based on reference milk samples with known contents of FA were used. These results indicate the usefulness of equations with R2cv greater than 95% in milk payment systems and the usefulness of equations with R2cv greater than 75% for animal breeding purposes.

"Global" and "local" predictions of dairy diet nutritional quality using near infrared reflectance spectroscopy.

The objective of the study was to evaluate performance of classic (global) and innovative (local) calibration techniques to monitor cattle diet, based on fecal near infrared reflectance spectroscopy (NIRS). A 3-yr on-farm survey (2005-2008) was carried out in Vietnam and La Reunion Island to collect animal, feed intake, and feces excretion data. Feed and feces were scanned by a Foss NIRsystem 5000 monochromator (Foss, Hillerød, Denmark) to estimate diet characteristics and nutrient digestibility. A data set including 1,322 diet-fecal pairs was built and used to perform global and local calibrations. Global equations gave satisfactory accuracy [coefficient of determination (R(2)) >0.8, 10% ≤ relative standard error of prediction (RSEP) ≤20%], whereas local equations gave good accuracy (R(2) >0.8, RSEP <10%) or excellent accuracy (R(2) >0.9, RSEP <10%) for the prediction of diet intake, quality, and digestibility. When validating the equations using the external individual data, both techniques were robust, with similar RSEP (8%) and R(2) (0.82) values. The predictive performance of global and local equations was improved (RSEP = 5% and R(2)=0.90) when averaged animal data from farm, visit, and similar milk production were used. In particular, local equations reduced RSEP by 43% and increased R(2) by 15%, on average, compared with those obtained from individual data. The low RSEP (4%), high R(2) (0.96), and good ratio performance deviation (RPD=5) illustrated the excellent accuracy and robustness of the local equations. Findings suggest the ability of fecal NIRS to successfully and more accurately predict diet properties (intake, quality, and digestibility) with local calibration techniques compared with classic global techniques, especially on an averaged data set. Local calibration techniques represent an alternative promising method and potentially a decision support tool to decide whether diets meet dairy cattle requirements or need to be modified.

Key parameters for the development of a NIR microscopic method for the quantification of processed by-products of animal origin in compound feedingstuffs.

The aim of this work is to show new advances in the analytical methods developed in the frame of the ban of processed animal by-products in compound feed that is currently applied within the European Union. With this aim, studies to develop a quantitative near infrared microscopy (NIRM) approach have been undertaken in order to fulfil future requirements of European legislation like the introduction of tolerance levels that would require for official control purposes the availability of specific quantitative methods. The capabilities of the NIRM method have been improved; no sample preparation is required and the acquisition parameters are optimised. Both the gross and the fine fractions of the samples are considered; the reflexion mode was used to analyse the gross raw fraction and the transmission mode was chosen to analyse the fine raw fraction. Parameters for reflexion analyses were already fixed in our previous studies while those of transmission mode have been determined in the present study. Because particles are too small, it is difficult to mark them; spectra were collected using the mapping technique. Quantitative analyses have been carried out for different percentages of adulteration (0.5, 1, 2 and 5%). Results were depending on the particle size distribution of the feed and of the fish meal which led to experimental values of adulteration varying between 0.13-0.92%, 0.93-3.7%, 2.42-5.83% and 1.95-9.39% for theoretical percentages of adulteration equal to 0.5, 1, 2 and 5%, respectively. The established protocol with the key parameters proposed has to be considered for the development of an accurate method of quantification.

Potential estimation of major mineral contents in cow milk using mid-infrared spectrometry.

Milk and dairy products are a major source of minerals, particularly calcium, involved in several metabolic functions in humans. Currently, several dairy products are enriched with calcium to prevent osteoporosis. The development of an inexpensive and fast quantitative analysis for minerals is required to offer dairy farmers an opportunity to improve the added value of the produced milk. The aim of this study was to develop 5 equations to measure Ca, K, Mg, Na, and P contents directly in bovine milk using mid-infrared (MIR) spectrometry. A total of 1,543 milk samples were collected between March 2005 and May 2006 from 478 cows during the Walloon milk recording and analyzed by MIR spectrometry. Using a principal component approach, 62 milk samples were selected by their spectral variability and separated in 2 calibration sets. Five outliers were detected and deleted. The mineral contents of the selected samples were measured by inductively coupled plasma atomic emission spectrometry. Using partial least squares combined with a repeatability file, 5 calibration equations were built to estimate the contents of Ca, K, Mg, Na, and P in milk. To assess the accuracy of the developed equations, a full cross-validation and an external validation were performed. The cross-validation coefficients of determination (R(2)cv) were 0.80, 0.70, and 0.79 for Ca, Na, and P, respectively (n = 57), and 0.23 and 0.50 for K and Mg, respectively (n = 31). Only Ca, Na, and P equations showed sufficient R(2)cv for a potential application. These equations were validated using 30 new milk samples. The validation coefficients of determination were 0.97, 0.14, and 0.88 for Ca, Na, and P, respectively, suggesting the potential to use the Ca and P calibration equations. The last 30 samples were added to the initial milk samples and the calibration equations were rebuilt. The R(2)cv for Ca, K, Mg, Na, and P were 0.87, 0.36, 0.65, 0.65, and 0.85, respectively, confirming the potential utilization of the Ca and P equations. Even if new samples should be added in the calibration set, the first results of this study showed the feasibility to quantify the calcium and phosphorus directly in bovine milk using MIR spectrometry.