Use and abuse of chemometrics in chromatography

Metabolite Fingerprinting Based on 1H-NMR Spectroscopy and Liquid Chromatography for the Authentication of Herbal Products

Herbal medicines (HMs) are regarded as one of the traditional medicines in health care to prevent and treat some diseases. Some herbal components such as turmeric and ginger are used as HMs, therefore the identification and confirmation of herbal use are very necessary. In addition, the adulteration practice, mainly motivated to gain economical profits, may occur by substituting the high price of HMs with lower-priced ones or by addition of certain chemical constituents known as Bahan Kimia Obat (chemical drug ingredients) in Indonesia. Some analytical methods based on spectroscopic and chromatographic methods are developed for the authenticity and confirmation of the HMs used. Some approaches are explored during HMs authentication including single-component analysis, fingerprinting profiles, and metabolomics studies. The absence of reference standards for certain chemical markers has led to exploring the fingerprinting approach as a tool for the authentication of HMs. During fingerprinting-based spectroscopic and chromatographic methods, the data obtained were big, therefore the use of chemometrics is a must. This review highlights the application of fingerprinting profiles using variables of spectral and chromatogram data for authentication in HMs. Indeed, some chemometrics techniques, mainly pattern recognition either unsupervised or supervised, were applied for this purpose.

Contribution to second-order calibration based on multivariate curve resolution with and without previous chromatographic synchronization

Rotational ambiguity in the bilinear solutions provided by multivariate curve resolution - alternating least-squares (MCR-ALS) leads to an additional source of uncertainty in the estimation of analyte concentrations by second-order multivariate calibration. The phenomenon is particularly important when measuring matrix instrumental data derived from liquid chromatography with spectral detection, where elution time profiles usually vary from sample to sample both in position and shape. This makes the data non-trilinear, precluding the use of unique trilinear decomposition models. The present report compares some analytical results achieved by: (1) the usual MCR-ALS analysis of augmented matrices built from raw matrix data and (2) a previously reported procedure based on synchronizing the MCR-ALS elution time profiles using correlation optimized warping (COW), reconstructing the augmented matrix with the spectra and the aligned chromatograms, and then applying MCR-ALS again with the trilinearity constraint, leading to unique solutions, which is possible because the warping process restores the trilinearity of the data. We show that alternative (2) does not solve the rotational ambiguity issues and artificially modifies the original data, without significant improvements in analytical performance. In the simulated systems, the best average errors for alternative (1) were about 2%, whereas for alternative (2) they were in the range 4-11%. For the experimental system, the corresponding errors were 2-3% and 3-4% respectively, i.e. with no significant improvement in going to alternative (2). All efforts should be directed to reduce the degree of rotational ambiguity by applying a full battery of chemically reasonable constraints.

Deep learning for vibrational spectral analysis: Recent progress and a practical guide

The development of chemometrics aims to provide an effective analysis approach for data generated by advanced analytical instruments. The success of existing analytical approaches in spectral analysis still relies on preprocessing and feature selection techniques to remove signal artifacts based on prior experiences. Data-driven deep learning analysis has been developed and successfully applied in many domains in the last few years. How to integrate deep learning with spectral analysis received increased attention for chemometrics. Approximately 20 recently published studies demonstrate that deep neural networks can learn critical patterns from raw spectra, which significantly reduces the demand for feature engineering. The composition of multiple processing layers improves the fitting and feature extraction capability and makes them applicable to various analytical tasks. This advance offers a new solution for chemometrics toward resolving challenges related to spectral data with rapidly increased sample numbers from various sources. We further provide a practical guide to the development of a deep convolutional neural network-based analytical workflow. The design of the network structure, tuning the hyperparameters in the training process, and repeatability of results is mainly discussed. Future studies are needed on interpretability and repeatability of the deep learning approach in spectral analysis.

Identification and Validation of Sensory-Active Compounds from Data-Driven Research: A Flavoromics Approach

In this study, highly predictive LC-MS features (retention time_ m/ z) derived from untargeted chemical fingerprinting-multivariate analysis (MVA) previously used to model flavor changes in citrus fruits related to aging (freshness) were further isolated and analyzed for sensory impact, followed by structural elucidation. The top 10 statistical features from two MVA approaches, partial least-squares data analysis (PLS-DA) and Random Forrest (RF), were purified to approximately 70% via multidimensional liquid chromatography-mass-directed fractionation to screen for sensory activity. When added to a 'fresh' orange flavor model system, 50-60% of the isolates were reported to cause a sensory change. From the subset of the actives identified, two compounds were selected, on the basis of statistical relevance, that were further purified to >97% for identification (MS, NMR) and for sensory descriptive analysis (DA). The compounds were identified as nomilin glucoside and a novel ionone glucoside. DA evaluation in the recombination orange model indicated both compounds statistically suppressed the perceived intensity of the "orange character" attribute, whereas the novel ionone glycoside also decreased the intensity of the floral character while increasing the green bean attribute intensity.

Discrimination of Polygoni Multiflori radix and Cynanchi Auriculati radix using ultra-high performance liquid chromatography fingerprints and chemical pattern recognition

In this work, a strategy was proposed to discriminate Polygoni Multiflori Radix (PMR) and its adulterant (Cynanchi Auriculati Radix, CAR). Ultra-high performance liquid chromatography (UHPLC) fingerprints were established to analyze samples containing PMR, CAR and mixtures simultaneously. Multivariate classification methods were applied to analyze the obtained UHPLC fingerprints, including principal component analysis (PCA), partial least square discriminant analysis (PLS-DA), soft independent modeling of class analogy (SIMCA), support vector machine discriminant analysis (SVMDA) and counter-propagation artificial neural network (CP-ANN). A plot of PCA score showed that PMR and CAR samples belonged to separate clusters (PMR class and CAR class), and samples of mixtures were located near PMR or CAR classes. Analysis by PLS-DA, SVMDA and CP-ANN performed well for recognition and prediction in terms of PMR and CAR samples. Moreover, the PLS-DA method performed best in the detection of adulterated samples, even if the adulterant was about 25%.

Combination of Liquid Chromatography with Multivariate Curve Resolution-Alternating Least-Squares (MCR-ALS) in the Quantitation of Polycyclic Aromatic Hydrocarbons Present in Paprika Samples

This work presents a strategy for quantitating polycyclic aromatic hydrocarbons (PAHs) in smoked paprika samples. For this, a liquid chromatographic method with fluorimetric detection (HPLC-FLD) was optimized. To resolve some interference co-eluting with the target analytes, the second-order multivariate curve resolution-alternating least-squares (MCR-ALS) algorithm has been employed combined with this liquid chromatographic method. Among the eight PAHs quantified (fluorene, phenanthrene, anthracene, pyrene, chrysene, benzo[a]anthracene, benzo[b]fluoranthene, and benzo[a]pyrene) by HPLC-FLD, only in the case of fluorene, pyrene, and benzo[b]fluoranthene was it necessary to apply the second-order algorithm for their resolution. Limits of detection and quantitation were between 0.015 and 0.45 mg/kg and between 0.15 and 1.5 mg/kg, respectively. Good recovery results (>80%) for paprika were obtained via the complete extraction procedure, consisting of an extraction from the matrix and the cleanup of the extract by means of silica cartridges. Higher concentrations of chrysene, benzo[a]anthracene, benzo[b]fluoranthene, and benzo[a]pyrene were found in the paprika samples, with respect to the maximal amounts allowed for other spices that are under European Regulation (EU) N° 2015/1933.

Checkpoints for Preliminary Identification of Small Molecules found Enriched in Autophagosomes and Activated Mast Cell Secretions Analyzed by Comparative UPLC/MSe

We report the use of ultra high performance liquid chromatography (UPLC) coupled with acquisition of low- and high-collision energy mass spectra (MSe) to explore small molecule compositions that are unique to either enriched-autophagosomes or secretions of chemically activated murine mast cells. Starting with thousands of features, each defined by a chromatographic retention time, m/z values and ion intensities, manual examination of the extracted ion chromatograms (XIC) of chemometrically selected features was essential to eliminate false positives, occurring at rates of 33, 14 and 37% in samples of three biological systems. Forty-six percent of features that passed the XIC-based checkpoint, had IDs in compound databases used here. From these, 19% of IDs had experimental high-collision energy MSe spectra that were in agreement with in-silico fragmentation. The importance of this second checkpoint was highligthed through validation with selected commercially available standards. This work illustrates that checkpoints in data processing are essential to ascertain reliability of unbiased metabolomic studies, thereby reducing the risk of generating 'false identifications' which are is a major concern as 'omics' data continue to proliferate and be used as platforms to lauch novel biological hypotheses.

Detection of discoloration in diesel fuel based on gas chromatographic fingerprints

In the countries of the European Community, diesel fuel samples are spiked with Solvent Yellow 124 and either Solvent Red 19 or Solvent Red 164. Their presence at a given concentration indicates the specific tax rate and determines the usage of fuel. The removal of these so-called excise duty components, which is known as fuel “laundering”, is an illegal action that causes a substantial loss in a government’s budget. The aim of our study was to prove that genuine diesel fuel samples and their counterfeit variants (obtained from a simulated sorption process) can be differentiated by using their gas chromatographic fingerprints that are registered with a flame ionization detector. To achieve this aim, a discriminant partial least squares analysis, PLS-DA, for the genuine and counterfeit oil fingerprints after a baseline correction and the alignment of peaks was constructed and validated. Uninformative variables elimination (UVE), variable importance in projection (VIP), and selectivity ratio (SR), which were coupled with a bootstrap procedure, were adapted in PLS-DA in order to limit the possibility of model overfitting. Several major chemical components within the regions that are relevant to the discriminant problem were suggested as being the most influential. We also found that the bootstrap variants of UVE-PLS-DA and SR-PLS-DA have excellent predictive abilities for a limited number of gas chromatographic features, 14 and 16, respectively. This conclusion was also supported by the unitary values that were obtained for the area under the receiver operating curve (AUC) independently for the model and test sets.

Comparative study of the alignment method on experimental and simulated chromatographic data

One of the major problems in the signal comparison of chromatographic data is the variability of response caused by instrumental drifts and others instabilities. Measures of quality control and evaluation of conformity are inherently sensitive to shift. It is essential to be able to compare test samples to reference samples in an evolutionary analytical environment by offsetting the inevitable drift. Therefore, prior to any multivariate analysis, the alignment of analytical signals is a compulsory preprocessing step. During recent years, many researchers have taken a greater interest in the study of the alignment. The present paper is an updated review on the alignment algorithms, methods, and improvements used in chromatography. The study is dedicated to one-dimensional signals. Several of the exposed methods have common theoretical bases and can differ through their optimization methods. The main issue for the operator is to choose the appropriate method according to the type of signals to be processed.

Modeling of the total antioxidant capacity of rooibos (Aspalathus linearis) tea infusions from chromatographic fingerprints and identification of potential antioxidant markers

Models to predict the total antioxidant capacity (TAC) of rooibos tea infusions from their chromatographic fingerprints and peak table data (content of individual phenolic compounds), obtained using HPLC with diode array detection, were developed in order to identify potential antioxidant markers. Peak table data included the content of 12 compounds, namely phenylpyruvic acid-2-O-glucoside, aspalathin, nothofagin, isoorientin, orientin, ferulic acid, quercetin-3-O-robinobioside, vitexin, hyperoside, rutin, isovitexin and isoquercitrin. The TAC values, measured using the oxygen radical absorbance capacity (ORAC) and DPPH radical scavenging assays, could be predicted from the peak table data or the chromatographic fingerprints (prediction errors 9-12%) using partial least squares (PLS) regression. Prediction models created from samples of only two production years could additionally be used to predict the TAC of samples from another production year (prediction errors<13%) indicating the robustness of the models in a quality control environment. Furthermore, the uninformative variable elimination (UVE)-PLS method was used to identify potential antioxidant markers for rooibos infusions. All individual phenolic compounds that were quantified were selected as informative variables, except vitexin, while UVE-PLS models developed from chromatographic fingerprints indicated additional antioxidant markers, namely (S)-eriodictyol-6-C-glucoside, (R)-eriodictyol-6-C-glucoside, aspalalinin and two unidentified compounds. The potential antioxidant markers should be validated prior to use in quality control of rooibos tea.

Morphological weighted penalized least squares for background correction

Backgrounds existing in the analytical signal always impair the effectiveness of signals and compromise selectivity and sensitivity of analytical methods. In order to perform further qualitative or quantitative analysis, the background should be corrected with a reasonable method. For this purpose, a new automatic method for background correction, which is based on morphological operations and weighted penalized least squares (MPLS), has been developed in this paper. It requires neither prior knowledge about the background nor an iteration procedure or manual selection of a suitable local minimum value. The method has been successfully applied to simulated datasets as well as experimental datasets from different instruments. The results show that the method is quite flexible and could handle different kinds of backgrounds. The proposed MPLS method is implemented and available as an open source package at http://code.google.com/p/mpls.

Targeted mass spectral ratio analysis: a new tool for gas chromatography-mass spectrometry

An algorithm, referred to as targeted mass spectral ratio analysis (TMSRA) is presented whereby the ratios between intensities as a function of mass channel (m/z) of a target analyte mass spectrum are used to automatically determine which m/z are sufficiently pure to quantify the analyte in a sample gas chromatogram. The standard perfluorotributylamine (PFTBA) was used to evaluate the reproducibility of the collected mass spectra, which aided in selecting a mass spectral threshold for TMSRA application to a subsequent case study. Results with PFTBA suggested that a threshold of all m/z at or above 1% of the highest recorded m/z intensity should be included for targeted analysis. For the case study, 1-heptene was selected as the target analyte and n-heptane was selected as the interfering compound. These two compounds were chosen since their mass spectra are very similar. Chromatographic data containing a pure peak for these analytes were extracted, and mathematically added at various temporal offsets to generate various degrees of chromatographic resolution, R(s), for the purpose of evaluating algorithm performance, and indeed, TMSRA successfully quantified 1-heptene. At the higher R(s) studied (0.6 ≤ R(s) ≤ 1.5) a deviation within ± 1% and a RSD generally below 1% were achieved for 1-heptene quantification. As the R(s) decreased, the deviation and RSD both increased. At a R(s)=0, a deviation of ≈ 9% and a RSD of ≈ 9% were achieved.

Enhanced chromatographic fingerprinting of herb materials by multi-wavelength selection and chemometrics

A strategy for multi-wavelength chromatographic fingerprinting of herbal materials, using high performance liquid chromatography with a UV-Vis diode array detector is presented. Valeriana officinalis was selected to show the proposed methodology since it is a widely used commercially available herbal drug, and because misfit with other valerian species is a current issue. The enhanced fingerprints were constructed by compiling into a single data vector the chromatograms from four wavelengths (226, 254, 280 and 326 nm), at which characteristic chemical constituents of studied herbs presented maximum absorbance. Chromatographic data pretreatment included baseline correction, normalization and correlation optimized warping. A simplex optimization was performed to retrieve the optimal values of the parameters used in the warping. General success rates of a classification above 90% were achieved by soft independent modeling of class analogy (SIMCA) and partial least squares discriminant analysis (PLS-DA). The sensitivity and specificity of constructed models were above 94%. Tests on laboratory-made mixtures showed that it is possible to detect adulterations or counterfeits with 5% foreign herbal material, even if it is from the Valerianaceae family. The results suggest that the proposed enhanced fingerprinting approach can be used to authenticate herb materials with complex chromatographic profiles.

Assessing principal component regression prediction of neurochemicals detected with fast-scan cyclic voltammetry

Principal component regression is a multivariate data analysis approach routinely used to predict neurochemical concentrations from in vivo fast-scan cyclic voltammetry measurements. This mathematical procedure can rapidly be employed with present day computer programming languages. Here, we evaluate several methods that can be used to evaluate and improve multivariate concentration determination. The cyclic voltammetric representation of the calculated regression vector is shown to be a valuable tool in determining whether the calculated multivariate model is chemically appropriate. The use of Cook's distance successfully identified outliers contained within in vivo fast-scan cyclic voltammetry training sets. This work also presents the first direct interpretation of a residual color plot and demonstrated the effect of peak shifts on predicted dopamine concentrations. Finally, separate analyses of smaller increments of a single continuous measurement could not be concatenated without substantial error in the predicted neurochemical concentrations due to electrode drift. Taken together, these tools allow for the construction of more robust multivariate calibration models and provide the first approach to assess the predictive ability of a procedure that is inherently impossible to validate because of the lack of in vivo standards.

Comparison of Several Methods of Chromatographic Baseline Removal with a New Approach Based on Quantile Regression

The article is intended to introduce and discuss a new quantile regression method for baseline detrending of chromatographic signals. It is compared with current methods based on polynomial fitting, spline fitting, LOESS, and Whittaker smoother, each with thresholding and reweighting approach. For curve flexibility selection in existing algorithms, a new method based on skewness of the residuals is successfully applied. The computational efficiency of all approaches is also discussed. The newly introduced methods could be preferred to visible better performance and short computational time. The other algorithms behave in comparable way, and polynomial regression can be here preferred due to short computational time.