Lars Nørgaard

Interval Partial Least-Squares Regression (iPLS): A Comparative Chemometric Study with an Example from Near-Infrared Spectroscopy

A new graphically oriented local modeling procedure called interval partial least-squares (iPLS) is presented for use on spectral data. The iPLS method is compared to full-spectrum partial least-squares and the variable selection methods principal variables (PV), forward stepwise selection (FSS), and recursively weighted regression (RWR). The methods are tested on a near-infrared (NIR) spectral data set recorded on 60 beer samples correlated to original extract concentration. The error of the full-spectrum correlation model between NIR and original extract concentration was reduced by a factor of 4 with the use of iPLS (r = 0.998, and root mean square error of prediction equal to 0.17% plato), and the graphic output contributed to the interpretation of the chemical system under observation. The other methods tested gave a comparable reduction in the prediction error but suffered from the interpretation advantage of the graphic interface. The intervals chosen by iPLS cover both the variables found by FSS and all possible combinations as well as the variables found by PV and RWR, and iPLS is still able to utilize the first-order advantage.

Chemometric Quantitation of the Active Substance (Containing C(N) in a Pharmaceutical Tablet Using Near-Infrared (NIR) Transmittance and NIR FT-Raman Spectra

In this study, near-infrared (NIR) transmittance and Raman spectroscopy chemometric calibrations of the active substance content of a pharmaceutical tablet were developed using partial least-squares regression (PLS). Although the active substance contained the strongly Raman active C≡N functional group, the best results were obtained with NIR transmittance, which highlights the difference between (microscopic) surface sampling and whole tablet diffuse transmittance sampling. The tablets exist in four dosages with only two different concentrations of active substance (5 mg (5.6% w/w), and 10, 15, and 20 mg (8.0% w/w) active substance per tablet). A calibration on all four dosages resulted in a prediction error expressed as the root mean squared error of cross-validation (RMSECV) of 0.30% w/w for the NIR transmittance calibration. The corresponding error when using Raman spectra was 0.56% w/w. Specially prepared calibration batches covering the range 85–115% of the nominal content for each dosage were added to the first sample set, and NIR transmittance calibrations on this set—containing coated as well as uncoated tablets—gave a further reduction in prediction errors to 0.21–0.289% w/w. This corresponds to relative prediction errors (RMSECV/ynom) of 2.6–3.7%. This is a reasonably low error when compared to the error of the chromatographic reference method, which was estimated to 3.5%.

Chemometrics in food science—a demonstration of the feasibility of a highly exploratory, inductive evaluation strategy of fundamental scientific significance

Abstract At the roots of science lies observation and data collection from the world as is and from which conclusions can be induced after classification. This is far from the present theory-driven, deductive, normative stage of science which depends heavily on modelling discrete functional factors in laboratory experiments and suppresses the aspect of interaction. In spite of its successes, science today has great difficulty in adapting to the changes which technology has created to cope with registering and evaluating real data from the world, such as in food production chains. This paper demonstrates that it is possible and profitable with the help of new technology to reintroduce an explorative, inductive strategy to investigate the chemistry of a complex food process as is with a minimum of a priori assumptions. The food process investigated is a sugar plant and the tools necessary in this strategy include a multivariate screening method (fluorescence spectroscopy), an arsenal of chemometric models (PCA, PLS, principal variables), including multiway models ( parafac , Tucker), and a computer. Not only can chemical criteria and process parameters throughout the process be validly predicted by the screening method, but process irregularities as well as chemical species can also be detected and validated by multiway chemometric techniques. Inspired by examples from the food area, the paper further discusses the nature of the exploration method in the selection of tools and data. The aim is to study complex processes as a whole in order to model interaction of the underlying latent functional factors which may later be defined more precisely by deductive methods. These methods in combination with an appropriate multivariate screening method allow for unique identification of objects—a significant prerequisite for a viable, exploratory, inductive data strategy which is needed as a fundamental complement to prevalent normative research in order to obtain a science on the interdisciplinary level.

Exploring the phenotypic expression of a regulatory proteome- altering gene by spectroscopy and chemometrics

Evaluating gene effects on proteomes and the resulting indirect pleiotropic effects through the cell machinery on the chemical phenotype constitutes a formidable challenge to the analytical chemist. This paper demonstrates that near-infrared (NIR) spectroscopy and chemometrics on the level of the barley seed phenotype is able to differentiate between genetic and environmental effects in a PCA model involving normal barley lines and the gene regulator lys3a in different genetic backgrounds. The gene drastically changes the proteome quantitatively and qualitatively, as displayed in two-dimensional electrophoresis, resulting in a radically changed amino acid and chemical composition. A synergy interval partial least squares regression model (si-PLSR) is tested to select combinations of spectral segments which have a high correlation to defined chemical components indicative of the lys3a gene, such as direct effects of the changed proteome, for example, the amide content, or indirect effects due to changes in carbohydrate and fat composition. It is concluded that the redundancy of biological information on the DNA sequence level is also represented at the phenotypic level in the dataset read by the NIR spectroscopic sensor from the chemical physical fingerprint. The PLS algorithm chooses spectral intervals which combine both direct and indirect proteome effects. This explains the robustness of NIR spectral predictions by PLSR for a wide range of chemical components. The new option of using spectroscopy, analytical chemistry and chemometrics in modeling the genetically based covariance of physical/chemical fingerprints of the intact phenotype in plant breeding and biotechnology is discussed.

Comparative vibrational spectroscopy for determination of quality parameters in amidated pectins as evaluated by chemometrics

Abstract The potential of vibrational spectroscopy and chemometrics as a reliable and fast method for the determination of important gel-forming parameters in amidated pectins has been investigated. For a set of 98 amidated pectin samples, six complete spectroscopic ensembles were recorded including NIR, FT-NIR, FT-IR and NIR FT-Raman spectroscopy. For each spectroscopic ensemble, quantitative models based on partial least squares regression (PLS) have been developed and compared. Chemometric models were constructed by dividing the spectroscopic ensembles up into a calibration set of 73 samples and an independent test set of 25 samples to evaluate the predictive ability of the models. The 98 amidated pectin samples span a degree of esterification (%DE) between 20 and 55% and a degree of amidation (%DA) between 4 and 24 per cent. From all six spectroscopic ensembles quantitative PLS models were obtained for %DE and %DA with RMSEP (root mean square error of prediction) ranging between 1.5 and 2.1 and between 1.1 and 2.1 for %DE and %DA respectively. In both cases the results are comparable to that of the experimental error of the quantitative chemical determination. Finally, different ways of selecting FT-IR spectral elements (variables) to correlate with %DE were compared. The method of principal variables (PV) was found to be superior to methods based on knowledge based selection and the resulting PV-based model was found to be slightly better than the PLS model using the full spectral information.

Prediction of technological quality (cooking loss and Napole Yield) of pork based on fresh meat characteristics

In order to investigate if cooking loss and Napole Yield can be predicted from various fresh meat characteristics, pH (1, 15, 30, 60, 120 min and 24 h post mortem), temperature (1, 15, 30, 60, 120 min and 24 h post mortem), water-holding capacity (Honikels drip loss method and centrifugation loss), and NMR T2 relaxation 24 h post mortem were measured in fresh porcine M. longissimus dorsi from 102 Hampshire crossbreeds of known RN(-) genotype. Subsequently, cooking loss and Napole Yield were determined on cooked and cured, cooked samples, respectively, and partial least squares regression (PLS) was carried out to investigate possible intercorrelations between the physico-chemical measurements performed on the fresh meat and cooking loss/Napole Yield. Significant correlations were found between NMR T2 relaxation measured in fresh pork 24 h post mortem and the cooking loss (R=0.64) and Napole Yield (R=0.58), whereas no correlations were found between traditionally applied methods such as pH measurements, Honikels method and centrifugation, and the cooking loss/Napole Yield. Consequently, it is concluded that NMR T2 relaxation characteristics of fresh pork in contrast to traditional fresh meat characteristics contain information about factors of importance for cooking loss/Napole Yield from cooked uncured/cured pork. The result implies that low-field (LF) NMR data from fresh meat reflects information about water compartmentalisation and mobility that is partly decisive for subsequent heat-induced changes of importance for the distribution of water within the cooked meat. In addition, the obtained results show that LF NMR data measured on fresh meat also seems to contain information about the inherent water of importance for the technological characteristics of the meat even when the meat is cured before cooking.

Non-linear calibration models for near infrared spectroscopy.

Different calibration techniques are available for spectroscopic applications that show nonlinear behavior. This comprehensive comparative study presents a comparison of different nonlinear calibration techniques: kernel PLS (KPLS), support vector machines (SVM), least-squares SVM (LS-SVM), relevance vector machines (RVM), Gaussian process regression (GPR), artificial neural network (ANN), and Bayesian ANN (BANN). In this comparison, partial least squares (PLS) regression is used as a linear benchmark, while the relationship of the methods is considered in terms of traditional calibration by ridge regression (RR). The performance of the different methods is demonstrated by their practical applications using three real-life near infrared (NIR) data sets. Different aspects of the various approaches including computational time, model interpretability, potential over-fitting using the non-linear models on linear problems, robustness to small or medium sample sets, and robustness to pre-processing, are discussed. The results suggest that GPR and BANN are powerful and promising methods for handling linear as well as nonlinear systems, even when the data sets are moderately small. The LS-SVM is also attractive due to its good predictive performance for both linear and nonlinear calibrations.

A multivariate chemometric approach to fluorescence spectroscopy

A multivariate approach to the solution of problems often encountered in the spectrofluorometry of natural samples, utilising information from whole spectra is presented. (a) Piecewise direct standardisation is implemented and employed to transfer emission spectra measured with two different xenon lamps of different ages as if the spectra were measured with the same lamp. (b) It has been shown using a multivariate analysis approach that it is possible to use the raw data points instead of the smoothed data based on an algorithm included in the instrument software by the manufacturer. (c) It is documented that Raman scattering does not hamper the performance of multivariate calibration; on the contrary, in an experiment with sugar samples the concentration prediction errors become about five times lower by including the whole emission spectrum in the analysis instead of using a univariate calibration based on an emission wavelength that only reflects the analyte of interest. (d) An algorithm for variable selection is implemented and employed in the selection of optimal excitation wavelengths. Among 13 emission spectra recorded for a sugar sample at different excitation wavelengths, four of these are chosen that describe 98.51% of the total variance in the original data. (e) Finally the combination of fluorescence spectroscopy and multivariate calibration with conventional chemical data according to the near-infrared black box model is presented. The refined sugar quality parameter, the ash content and the fluorescence emission spectra are correlated by a partial least-squares regression model. Five experiments employing different monochromator slit widths and sugar concentrations are performed, and the best correlation obtained by full cross-validation of the 15 sugar samples is R = 0.98.

Ghanaian cocoa bean fermentation characterized by spectroscopic and chromatographic methods and chemometrics.

Export of cocoa beans is of great economic importance in Ghana and several other tropical countries. Raw cocoa has an astringent, unpleasant taste, and flavor, and has to be fermented, dried, and roasted to obtain the characteristic cocoa flavor and taste. In an attempt to obtain a deeper understanding of the changes in the cocoa beans during fermentation and investigate the possibility of future development of objective methods for assessing the degree of fermentation, a novel combination of methods including cut test, colorimetry, fluorescence spectroscopy, NIR spectroscopy, and GC-MS evaluated by chemometric methods was used to examine cocoa beans sampled at different durations of fermentation and samples representing fully fermented and dried beans from all cocoa growing regions of Ghana. Using colorimetry it was found that samples moved towards higher a* and b* values as fermentation progressed. Furthermore, the degree of fermentation could, in general, be well described by the spectroscopic methods used. In addition, it was possible to link analysis of volatile compounds with predictions of fermentation time. Fermented and dried cocoa beans from the Volta and the Western regions clustered separately in the score plots based on colorimetric, fluorescence, NIR, and GC-MS indicating regional differences in the composition of Ghanaian cocoa beans. The study demonstrates the potential of colorimetry and spectroscopic methods as valuable tools for determining the fermentation degree of cocoa beans. Using GC-MS it was possible to demonstrate the formation of several important aroma compounds such 2-phenylethyl acetate, propionic acid, and acetoin and the breakdown of others like diacetyl during fermentation. Practical Application: The present study demonstrates the potential of using colorimetry and spectroscopic methods as objective methods for determining cocoa bean quality along the processing chain. Development of objective methods for determining cocoa bean quality will be of great importance for quality insurance within the fields of cocoa processing and raw material control in chocolate producing companies.

Optimization of flow-injection systems for determination of substrates by means of enzyme amplification reactions and chemiluminescence detection.

A flow-injection system is described that incorporates a small column reactor containing two co-immobilized, synergistically operating oxidoreductases, allowing determination of minute amounts of substrates by means of enzyme amplification and subsequent chemiluminescence detection of the hydrogen peroxide generated in the repeated redox cycling. With lactate oxidase and lactate dehydrogenase, and taking advantage of the fact that the enzymatic degradation step and the ensuing detection step can be individually optimized, the FIA-system has been optimized by factorial experiments to yield an amplification factor of over 140 for each of the two substrates lactate and pyruvate. With a linear calibration range of 0-6muM, the limits of detection for the two species were 48 and 103nM, respectively, and the sampling rate was 50-60/hr. The optimized system has also been employed for assay of glucose by utilizing a column reactor with immobilized glucose oxidase and glucose dehydrogenase, but yielded amplification factors of only 3-4. The large discrepancy in the performance of the two enzyme systems is discussed.