Benchtop volatilomics supercharged: How machine learning based design of experiment helps optimizing untargeted GC-IMS gas phase metabolomics.

Gas chromatography-ion mobility spectrometry (GC-IMS) plays a significant role in both targeted and non-targeted analyses. However, the non-linear behavior of IMS and its complex ion chemistry pose challenges for finding optimal experimental conditions using existing methodologies. To address these issues, integrating machine learning (ML) strategies offers a promising approach. In this study, we propose a hybrid strategy, combining design of experiment (DOE) and machine learning (ML) for optimizing GC-IMS conditions in non-targeted volatilomic/flavoromic analysis, with saffron volatiles as a case study. To begin, a rotatable circumscribed central composite design (CCD) is used to define five influential GC-IMS factors of sample amount, headspace temperature, incubation time, injection volume, and split ratio. Subsequently, two ML models are utilized: multiple linear regression (MLR) as a linear model and Bayesian regularized-artificial neural network (BR-ANN) as a nonlinear model. These models are employed to predict the response variables of total peak areas (PAs) and the number of detected peaks (PNs) in GC-IMS. The findings show that there is a direct correlation between the factors in GC-IMS and the PNs, suggesting that MLR is a suitable approach for building a model in this scenario. However, the PAs exhibit nonlinear behavior, suggesting that BR-ANN is better suitable to capture this complexity. Notably, Derringer's desirability function is utilized to integrate the PAs and PNs, and in this scenario, MLR demonstrates satisfactory performance in modeling the GC-IMS factors.

Automated 2D peak detection in gas chromatography-ion mobility spectrometry through persistent homology.

BACKGROUND: Gas chromatography-ion mobility spectrometry (GC-IMS) is a powerful analytical technique which has gained widespread use in a variety of fields. Detecting peaks in GC-IMS data is essential for chemical identification. Topological data analysis (TDA) has the ability to record alterations in topology throughout the entire spectrum of GC-IMS data and retain this data in diagrams known as persistence diagrams. To put it differently, TDA naturally identifies characteristics such as mountains, volcanoes, and their higher-dimensional equivalents within the original data and measures their significance. RESULTS: In the present contribution, a novel approach based on persistent homology (a flagship technique of TDA) is suggested for automatic 2D peak detection in GC-IMS. For this purpose, two different GC-IMS data examples (urine and olive oil) are used to show the performance of the proposed method. The outputs of the algorithm are GC-IMS chromatogram with detected peaks, persistence plot showing the importance (intensity) of the detected peaks and a table with retention times (RT), drift times (DT), and persistence scores of detected peaks. The RT and DT can be used for identification of the peaks and persistence scores for quantitation. Additionally, watershed segmentation is applied to the GC-IMS images to index individual peaks and segment overlapping compounds allowing for a more accurate identification and quantification of individual peaks. SIGNIFICANCE: Inspection of the results for GC-IMS datasets showed the accurate and reliable performance of the proposed strategy based on persistent homology for automatic 2D GC-IMS peak detection for qualitative and quantitative analysis. In addition, this approach can be easily extended to other types of hyphenated chromatographic and/or spectroscopic data.

Comparison of a portable Vis-NIR hyperspectral imaging and a snapscan SWIR hyperspectral imaging for evaluation of meat authenticity.

The performance of visible-near infrared hyperspectral imaging (Vis-NIR-HSI) (400-1000 nm) and shortwave infrared hyperspectral imaging (SWIR-HSI) (1116-1670 nm) combined with different classification and regression (linear and non-linear) multivariate methods were assessed for meat authentication. In Vis-NIR-HSI, total accuracies in the prediction set for SVM and ANN-BPN (the best classification models) were 96 and 94 % surpassing the performance of SWIR-HSI with 88 and 89 % accuracy, respectively. In Vis-NIR-HSI, the best-obtained coefficient of determinations for the prediction set (R2p) were 0.99, 0.88, and 0.99 with root mean square error in prediction (RMSEP) of 9, 24 and 4 (%w/w) for pork in beef, pork in lamb and pork in chicken, respectively. In SWIR-HSI, the best-obtained R2p were 0.86, 0.77, and 0.89 with RMSEP of 16, 23 and 15 (%w/w) for pork in beef, pork in lamb and pork in chicken, respectively. The results ascertain that Vis-NIR-HSI coupled with multivariate data analysis has better performance rather than SWIR-HIS.

Quantitative Mass Spectrometry Imaging Using Multivariate Curve Resolution and Deep Learning: A Case Study.

In the present contribution, a novel approach based on multivariate curve resolution and deep learning (DL) is proposed for quantitative mass spectrometry imaging (MSI) as a potent technique for identifying different compounds and creating their distribution maps in biological tissues without need for sample preparation. As a case study, chlordecone as a carcinogenic pesticide was quantitatively determined in mouse liver using matrix-assisted laser desorption ionization-MSI (MALDI-MSI). For this purpose, data from seven standard spots containing 0 to 20 picomoles of chlordecone and four unknown tissues from the mouse livers infected with chlordecone for 1, 5, and 10 days were analyzed using a convolutional neural network (CNN). To solve the lack of sufficient data for CNN model training, each pixel was considered as a sample, the designed CNN models were trained by pixels in training sets, and their corresponding amounts of chlordecone were obtained by multivariate curve resolution-alternating least-squares (MCR-ALS). The trained models were then externally evaluated using calibration pixels in test sets for 1, 5, and 10 days of exposure, respectively. Prediction R2 for all three data sets ranged from 0.93 to 0.96, which was superior to support vector machine (SVM) and partial least-squares (PLS). The trained CNN models were finally used to predict the amount of chlordecone in mouse liver tissues, and their results were compared with MALDI-MSI and GC-MS methods, which were comparable. Inspection of the results confirmed the validity of the proposed method.

Mutual information map as a new way for exploring the independence of chemically meaningful solutions in two-component analytical data.

In the present contribution, a new approach based on mutual information (MI) is proposed for exploring the independence of feasible solutions in two component systems. Investigating how independent are different feasible solutions can be a way to bridge the gap between independent component analysis (ICA) and multivariate curve resolution (MCR) approaches and, to the best of our knowledge, has not been investigated before. For this purpose, different chromatographic and hyperspectral imaging (HSI) datasets were simulated, considering different noise levels and different degrees of overlap for two-component systems. Feasible solutions were then calculated by both grid search (GS) and Lawton-Sylvester (LS) plots. MI map which is the plot of MI vs. rotation matrix elements was used to estimate the degree of independence between different solutions. Inspection of the results showed that the different solutions in the feasible bands correspond to different MI values and that those values are lower for spectral profiles (more independent) than for concentration profiles (more dependent) as expected from the duality concept and the opposite is true. In addition, component profiles are found near more dependent solutions for concentration profiles and near less dependent solutions for spectral profiles which is due to the fact that "independence" constraint was applied to the spectral profiles in ICA algorithms. The performance of three well-known ICA algorithms (mean-field independent component analysis (MF-ICA), mutual information-based least dependent component analysis (MILCA) and joint approximate diagonalization of eigenmatrices (JADE)) as well as MCR-alternating least squares (MCR-ALS) were investigated. MI maps showed that the solutions of MF-ICA and MCR-ALS are in the feasible bands but the MILCA and JADE solutions which are just based on the independence maximization are outside the MI maps.

Vis-NIR hyperspectral imaging coupled with independent component analysis for saffron authentication.

In the present contribution, visible-near infrared hyperspectral imaging (Vis-NIR-HSI) combined with a novel chemometric approach based on mean-filed independent component analysis (MF-ICA) followed by multivariate classification techniques is proposed for saffron authentication and adulteration detection. First, MF-ICA was used to exploit pure spatial and spectral profiles of the components. Then, principal component analysis (PCA) and hierarchical cluster analysis (HCA) were used to find patterns of authentic samples based on their distribution maps. Then, detection of five common plant-derived adulterants of saffron including safflower, saffron style, calendula, rubia and turmeric were investigated. For this purpose, partial least squares-discriminant analysis (PLS-DA) for supervised classification to find a boundary between authentic and adulterated saffron samples. Classification accuracies for all models for calibration and prediction sets were 100 %. Finally, a mixed dataset was prepared and analyzed using the proposed strategy which again 100 % of accuracies for calibration and prediction sets were obtained. At the end, data driven soft independent modelling of class analogy (dd-SIMCA) was used to evaluate model for class modeling. Sensitivity was 95% for authentic class and specificities for all adulterants were 100%.

Sensitivity and generalized analytical sensitivity expressions for quantitative analysis using convolutional neural networks.

In recent years, convolutional neural networks and deep neural networks have been used extensively in various fields of analytical chemistry. The use of these models for calibration tasks has been highly effective; however, few reports have been published on their properties and characteristics of analytical figures of merit. Currently, most performance measures for these types of networks only incorporate some function of prediction error. While useful, these measures are incomplete and cannot be used as an objective comparison among different models. In this report, a new method for calculating the sensitivity of any type of neural network is proposed and studied on both simulated and real datasets. Generalized analytical sensitivity is defined and calculated for neural networks as an additional figure of merit. Moreover, the dependence of convolutional neural networks on regularization dataset size is studied and compared with other conventional calibration methods.

Big (Bio)Chemical Data Mining Using Chemometric Methods: A Need for Chemists.

This Review summarizes how big (bio)chemical data (BBCD) can be analyzed with multivariate chemometric methods and highlights some of the important challenges faced by modern analytical researches. Here, the potential of chemometric methods to solve BBCD problems that are being encountered in chromatographic, spectroscopic and hyperspectral imaging measurements will be discussed, with an emphasis on their applications to omics sciences. In addition, insights and perspectives on how to address the analysis of BBCD are provided along with a discussion of the procedures necessary to obtain more reliable qualitative and quantitative results. In this Review, the importance of "big data" and of their relevance to (bio)chemistry are first discussed. Thereafter, analytical tools which can produce BBCD are presented as well as the theoretical background of chemometric methods and their limitations when they are applied to BBCD. Finally, the importance of chemometric methods for the analysis of BBCD in different chemical disciplines is highlighted with some examples. In this work, we have tried to cover many of the current applications of big data analysis in the (bio)chemistry field.

Chemometrics-assisted isotope ratio fingerprinting based on gas chromatography/combustion/isotope ratio mass spectrometry for saffron authentication.

In the present contribution, the capability of isotopic ratio mass spectrometry (IRMS) for saffron authentication and detection of four common plant-derived adulterants (marigold flower, safflower, rubia, and saffron style) was investigated. For this purpose, 62 authentic saffron samples were analyzed by elemental analyzer-IRMS (EA-IRMS) and gas chromatography-combustion-IRMS (GC-C-IRMS). In this regard, EA-IRMS and GC-C-IRMS isotope fingerprints of carbon-13 and nitrogen-15 isotopes of saffron components were provided and then analyzed by chemometric methods. Principal component analysis (PCA) showed two different behaviors regarding two main regions. Then, a representative saffron sample was provided to study adulteration. On this matter, binary mixtures of saffron and adulterants were prepared at five different weight percentages (5%, 10%, 15%, 25%, and 35%) and analyzed by EA-IRMS and GC-C-IRMS. Data-driven soft independent modeling of class analogy (DD-SIMCA) was used to model authentic saffron samples and find a boundary between authentic and adulterated samples with a sensitivity of 100% by GC-C-IRMS. After that, discriminant models of linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), and partial least squares-discriminant analysis (PLS-DA) were tested to find the best discrimination line and also detection of the lowest level of adulterants. Among different models, the QDA model outperformed other methods and showed the ability to predict adulterants at 5% w/w level with 100% accuracy and precision. Finally, the developed QDA model was successfully used to discriminate a set of mixed samples of saffron and four adulterants as well as some commercial samples.

External parameter orthogonalization-support vector machine for processing of attenuated total reflectance-mid-infrared spectra: A solution for saffron authenticity problem.

In the present work, a new approach based on external parameter orthogonalization combined with support vector machine (EPO-SVM) is proposed for processing of attenuated total reflectance-Fourier transform mid-infrared (ATR-FT-MIR) spectra with the goal of solving authentication problem in saffron, the most expensive spice in the world. First, one-hundred authentic saffron samples are clustered by principal component analysis (PCA) with EPO as the best preprocessing strategy. Then, EPO-SVM is used for the detection of four commonly used plant-derived adulterants (i.e. safflower, calendula, rubia, and style) in binary mixtures (saffron and each of plant adulterants) and its performance is compared with other common classification methods. The obtained results showed that the EPO-SVM approach has a much better classification accuracy (>95%) than other methods (accuracy<89.2%). Finally, two different sample sets including mixture of saffron and four plant adulterants and commercial saffron samples are used for validation of the developed EPO-SVM model. In this regard, classification figures of merit in terms of sensitivity, specificity and accuracy were respectively 96.6%, 97.1%, and 96.8% which showed good classification performance. It is concluded that the proposed EPO-PCA and EPO-SVM approaches can be considered as reliable tools for authentication and adulteration detection in saffron samples.

The Feasibility of Two Handheld Spectrometers for Meat Speciation Combined with Chemometric Methods and Its Application for Halal Certification.

Handheld visible-near-infrared (Vis-NIR) and near-infrared (NIR) spectroscopy can be cost-effective, rapid, non-destructive and transportable techniques for identifying meat species and may be valuable for enforcement authorities, retail and consumers. In this study, a handheld Vis-NIR (400-1000 nm) and a handheld NIR (900-1700 nm) spectrometer were applied to discriminate halal meat species from pork (halal certification), as well as speciation of intact and ground lamb, beef, chicken and pork (160 meat samples). Several types of class modeling multivariate approaches were applied. The presented one-class classification (OCC) approach, especially with the Vis-NIR sensor (95-100% correct classification rate), was found to be suitable for the application of halal from non-halal meat-species discrimination. In a discriminant approach, using the Vis-NIR data and support vector machine (SVM) classification, the four meat species tested could be classified with accuracies of 93.4% and 94.7% for ground and intact meat, respectively, while with partial least-squares discriminant analysis (PLS-DA), classification accuracies were 87.4% (ground) and 88.6% (intact). Using the NIR sensor, total accuracies of the SVM models were 88.2% and 81.5% for ground and intact meats, respectively, and PLS-DA classification accuracies were 88.3% (ground) and 80% (intact). We conclude that the Vis-NIR sensor was most successful in the halal certification (OCC approaches) and speciation (discriminant approaches) for both intact and ground meat using SVM.

Comparison of near-infrared (NIR) and mid-infrared (MIR) spectroscopy based on chemometrics for saffron authentication and adulteration detection.

In this work, the potential of near-infrared (NIR) and mid-infrared (MIR) spectroscopy along with chemometrics was investigated for authentication and adulteration detection of Iranian saffron samples. First, authentication of one-hundred saffron samples was examined by principal component analysis (PCA). The results showed the NIR spectroscopy can better predict the origin of samples than the MIR. Next, partial least squares-discriminant analysis (PLS-DA) was developed to detect four common plant-derived adulterants (i.e., saffron style, calendula, safflower, and rubia). In all cases, PLS-DA classification figures of merit in terms of sensitivity, specificity, error rate and accuracy were satisfactory for both NIR and MIR datasets. The built models were then successfully validated using test set and also commercial samples. Finally, partial least squares regression (PLSR) was used to estimate the amount of adulteration. In this case, only NIR showed a good performance with regression coefficients (R2) in range of 0.95-0.99.

Chemometric techniques coupled with NMR for matabolic profiling of lettuce exposed to polycyclic aromatic hydrocarbones.

Treated waste water (TWW) quality varies due to the occurrence of polycyclic aromatic hydrocarbons (PAHs) up to low µg L-1. In this study, a non-targeted metabolomic analysis was performed on lettuce (Lactuca sativa L) exposed to 4 PAHs by irrigation. The plants were watered with different concentrations of contaminants (0-100 µg L-1) for 39 days under controlled conditions and then harvested, extracted and analyzed by nuclear magnetic resonance (NMR). Different chemometric tools based on principal component analysis (PCA) and partial least square discriminant analysis (PLS-DA) are proposed for the analysis of the complex data sets generated in the different exposure experiments. Furthermore, Analysis of Variance Simultaneous Component Analysis (ASCA) of changes on metabolite peaks showed significant PAHs concentration and exposure time-dependent changes, clearly differentiating between exposed and non-exposed samples.

Combining multivariate image analysis with high-performance thin-layer chromatography for development of a reliable tool for saffron authentication and adulteration detection.

In this work, high-performance thin-layer chromatography (HPTLC) coupled with multivariate image analysis (MIA) is proposed as a fast and reliable tool for authentication and adulteration detection of Iranian saffron samples based on their HPTLC fingerprints. At first, the secondary metabolites of saffron were extracted using ultrasonic-assisted solvent extraction (UASE) which was optimized using central composite design (CCD). Next, the RGB coordinates of HPTLC images were used for estimation of saffron origin based on principal component analysis (PCA). The PCA scores plot showed that saffron samples were clustered into two clear-cut groups which was 92% matched with the geographical origins of the samples. In the next step, common plant-derived adulterants of saffron including safflower, saffron style, calendula, and rubia were investigated with MIA analysis of HPTLC images using partial least squares-discriminant analysis (PLS-DA) at 5-35% (w/w) levels. The PLS-DA results showed proper classification of saffron and adulterants with sensitivity 99.14%, specificity 96.94%, error rate 1.96% and accuracy 98.04. Also, the effect of HPTLC injection volume on the performance of the proposed strategy was evaluated. The ability of the proposed method was then investigated by analyzing two additional sample sets including mixed samples of four plant-derived adulterants and adulterated commercial samples. Sensitivity and specificity of this model were 100% which confirmed its validity.

Dataset of the application of handheld NIR and machine learning for chicken fillet authenticity study.

Diffuse reflectance near-infrared (NIR) data (908-1676 nm) of chicken breast fillets was recorded in a non-destructive way using a portable miniaturised NIR spectrometer. The NIR data was used to discriminate between fresh and thawed breast fillets and to determine the birds' growth conditions. NIR data was recorded of 153 commercial supermarket chicken fillet samples by applying the NIR device equipped with the standard issue collar on the samples in three different ways: (i) directly on the meat (ii) through the top foil of the package (i.e. with an air pocket between the foil and the breast fillet), and (iii) through the top foil with the packaging turned bottom up (i.e. no air pocket between the foil and the breast fillet). In order to generate thawed samples, the fresh samples were frozen and subsequently thawed. The freshness of the fillets was checked using ß-hydroxyacyl-CoA-dehydrogenase of 13% of the sample set. Five NIR spectra were collected per measurement mode from each sample resulting in 4590 raw NIR spectra. Multivariate statistics was applied and the interpretation of these calculations can be found in Parastar et al. [1]. The NIR data has a reuse potential for follow-up studies of chicken breast fillet authentication using a similar brand NIR device or to serve as calibration transfer data.

Analytical Figures of Merit for Multisensor Arrays.

Multisensor arrays employing various sensing principles are a rapidly developing field of research as they allow simple and inexpensive quantification of various parameters in complex samples. Quantitative analysis with such systems is based on multivariate regression techniques, and deriving of traditional analytical figures of merit (e.g., sensitivity, selectivity, limit of detection, and limit of quantitation) for such systems is not obvious and straightforward. Nevertheless, it is absolutely needed for further development of the multisensor research field and for introducing these instruments into the general context of analytical chemistry. Here, we report on the protocol for calculation of sensitivity, selectivity, and detection limits for multisensor arrays. The results are provided and discussed in detail for several real-world data sets.

Development of multi-response optimization and quadratic calibration curve for determination of ten pesticides in complex sample matrices using QuEChERS dispersive liquid-liquid microextraction followed by gas chromatography.

In this study, QuEChERS combined with dispersive liquid-liquid microextraction is developed for extraction of ten pesticides in complex sample matrices of water and milk. In this regard, effective factors of proposed extraction technique combined with gas chromatography with flame ionization detector were designed, modeled, and optimized using central composite design, multiple linear regression, and Nelder-Mead simplex optimization. Later, univariate calibration model for ten pesticides was developed in concentration range of 0.5-100 ng/mL. Surprisingly, quadratic calibration behavior was observed for some of the pesticides. In this regard, Mandel's test was used for evaluating linearity and types of calibration equation. Finally, four pesticides followed linear calibration curve with sensitivity (0.23-0.66 mL/ng), analytical sensitivity (0.20-0.32), regression coefficient (0.988-0.995), limit of detection (0.39-1.83 ng/mL), and limit of quantitation (1.30-6.10 ng/mL) and six of them followed quadratic calibration curve with sensitivity (0.18-0.93 mL/ng), analytical sensitivity (0.25-0.86), regression coefficient (0.944-0.999), limit of detection (0.59-1.92 ng/mL), and limit of quantitation (1.96-6.40 ng/mL). The calculated limits of detection were below the maximum residue limits according to European Union pesticides database of European Commission. Finally, the proposed analytical method was used for determination of ten pesticides in water and milk samples.

Gas chromatographic fingerprint analysis of secondary metabolites of Stachys lanata (Stachys byzantine C. Koch) combined with antioxidant activity modelling using multivariate chemometric methods.

S. lanata has been traditionally used as a medicinal plant due to its various biological activities such as antioxidant activity. Therefore, identification and quality control studies of this plant are of great importance. To this end, gas chromatography (GC) combined with chemometrics was proposed for fingerprint analysis of S. lanata samples. This study sought to classify GC fingerprints of twenty-eight S. lanata samples from eight different regions of Iran and more importantly, to correlate fingerprints to the antioxidant activity to select S. lanata volatile antioxidant markers. S. lanata samples were classified into five and three classes using partial least squares-discriminant analysis (PLS-DA) according to their GC fingerprints and antioxidant peaks, respectively. The results of PLS regression (PLS-R) and variable importance in projection (VIP) showed that phenol, 2,4-bis (1,1-dimethylethyl)-, hexadecanoic acid- ethyl ester, vitamin E, Beta- sitosterol, and 1- monolinoleoylglycerol trimethylsily ether are volatile antioxidant markers of S. lanata samples.

Quantitative analysis of multiple high-resolution mass spectrometry images using chemometric methods: quantitation of chlordecone in mouse liver.

In this work, a chemometrics-based strategy is developed for quantitative mass spectrometry imaging (MSI). In this regard, quantification of chlordecone as a carcinogenic organochlorinated pesticide (C10Cll0O) in mouse liver using the matrix-assisted laser desorption ionization MSI (MALDI-MSI) method is used as a case study. The MSI datasets corresponded to 1, 5 and 10 days of mouse exposure to the standard chlordecone in the quantity range of 0 to 450 µg g-1. The binning approach in the m/z direction is used to group high resolution m/z values and to reduce the big data size. To consider the effect of bin size on the quality of results, three different bin sizes of 0.25, 0.5 and 1.0 were chosen. Afterwards, three-way MSI data arrays (two spatial and one m/z dimensions) for seven standards and four unknown samples were column-wise augmented with m/z values as the common mode. Then, these datasets were analyzed using multivariate curve resolution-alternating least squares (MCR-ALS) using proper constraints. The resolved mass spectra were used for identification of chlordecone in the presence of a complex background and interference. Additionally, the augmented spatial profiles were post-processed and 2D images for each component were obtained in calibration and unknown samples. The sum of these profiles was utilized to set the calibration curve and to obtain the analytical figures of merit (AFOMs). Inspection of the results showed that the lower bin size (i.e., 0.25) provides more accurate results. Finally, the obtained results by MCR for three datasets were compared with those of gas chromatography-mass spectrometry (GC-MS) and MALDI-MSI. The results showed that the MCR-assisted method gives a higher amount of chlordecone than MALDI-MSI and a lower amount than GC-MS. It is concluded that a combination of chemometric methods with MSI can be considered as an alternative way for MSI quantification.

Chemometrics comparison of gas chromatography with mass spectrometry and comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry Daphnia magna metabolic profiles exposed to salinity.

The performances of gas chromatography with mass spectrometry and of comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry are examined through the comparison of Daphnia magna metabolic profiles. Gas chromatography with mass spectrometry and comprehensive two-dimensional gas chromatography with mass spectrometry were used to compare the concentration changes of metabolites under saline conditions. In this regard, a chemometric strategy based on wavelet compression and multivariate curve resolution-alternating least squares is used to compare the performances of gas chromatography with mass spectrometry and comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry for the untargeted metabolic profiling of Daphnia magna in control and salinity-exposed samples. Examination of the results confirmed the outperformance of comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry over gas chromatography with mass spectrometry for the detection of metabolites in D. magna samples. The peak areas of multivariate curve resolution-alternating least squares resolved elution profiles in every sample analyzed by comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry were arranged in a new data matrix that was then modeled by partial least squares discriminant analysis. The control and salt-exposed daphnids samples were discriminated and the most relevant metabolites were estimated using variable importance in projection and selectivity ratio values. Salinity de-regulated 18 metabolites from metabolic pathways involved in protein translation, transmembrane cell transport, carbon metabolism, secondary metabolism, glycolysis, and osmoregulation.