Partial Least Square Regression Discriminant Analysis Research Articles

Identification of Metabolic Biomarkers in Relation to Methotrexate Response in Early Rheumatoid Arthritis.

This study aimed to identify baseline metabolic biomarkers for response to methotrexate (MTX) therapy in rheumatoid arthritis (RA) using an untargeted method. In total, 82 baseline plasma samples (41 insufficient responders and 41 sufficient responders to MTX) were selected from the Treatment in the Rotterdam Early Arthritis Cohort (tREACH, trial number: ISRCTN26791028) based on patients’ EULAR response at 3 months. Metabolites were assessed using high-performance liquid chromatography-quadrupole time of flight mass spectrometry. Differences in metabolite concentrations between insufficient and sufficient responders were assessed using partial least square regression discriminant analysis (PLS-DA) and Welch’s t-test. The predictive performance of the most significant findings was assessed in a receiver operating characteristic plot with area under the curve (AUC), sensitivity and specificity. Finally, overrepresentation analysis was performed to assess if the best discriminating metabolites were enriched in specific metabolic events. Baseline concentrations of homocystine, taurine, adenosine triphosphate, guanosine diphosphate and uric acid were significantly lower in plasma of insufficient responders versus sufficient responders, while glycolytic intermediates 1,3-/2,3-diphosphoglyceric acid, glycerol-3-phosphate and phosphoenolpyruvate were significantly higher in insufficient responders. Homocystine, glycerol-3-phosphate and 1,3-/2,3-diphosphoglyceric acid were independent predictors and together showed a high AUC of 0.81 (95% CI: 0.72–0.91) for the prediction of insufficient response, with corresponding sensitivity of 0.78 and specificity of 0.76. The Warburg effect, glycolysis and amino acid metabolism were identified as underlying metabolic events playing a role in clinical response to MTX in early RA. New metabolites and potential underlying metabolic events correlating with MTX response in early RA were identified, which warrant validation in external cohorts.

Near-infrared spectroscopy as a new method for post-harvest monitoring of white truffles

The quality of truffles is related to their maturation stage, and good maturation degree is important in order to have a truffle with valuable organoleptic properties. An innovative, rapid and reliable technique to determine the maturation degree of truffles and monitor their post-harvest ageing process is proposed. Near-infrared (NIR) spectroscopy was used for the first time to monitor the post-harvest ripening of white truffles of two different species (Tuber magnatum and T. borchii). Optical microscopy was flanked by NIR spectroscopy imaging to monitor fruiting bodies at different degrees of maturation. The NIR profile of truffles provided statistical differentiation between the biochemical composition of different tissues (vegetative hyphae and ascospores). The optical evaluation of the degree of maturation was correlated to spectral information in order to obtain a non-operator dependent, rapid, simple and cost-effective method to evaluate the degree of maturation of white truffles. Partial least square regression (PLS) was used for multivariate calibration of NIR spectra against maturation degree, and ripening curves were obtained (with correlation coefficients R2 = 0.95 for T. magnatum and R2 = 0.76 for T. borchii). Moreover, a classification method based on PLS discriminant analysis of truffle in three maturity stages was developed, achieving a total of correct classification rate of 83%.

A deep learning- and partial least square regression-based model observer for a low-contrast lesion detection task in CT.

This work aims to develop a new framework of image quality assessment using deep learning-based model observer (DL-MO) and to validate it in a low-contrast lesion detection task that involves CT images with patient anatomical background. The DL-MO was developed using the transfer learning strategy to incorporate a pretrained deep convolutional neural network (CNN), a partial least square regression discriminant analysis (PLS-DA) model and an internal noise component. The CNN was previously trained to achieve the state-of-the-art classification accuracy over a natural image database. The earlier layers of the CNN were used as a deep feature extractor, with the assumption that similarity exists between the CNN and the human visual system. The PLSR model was used to further engineer the deep feature for the lesion detection task in CT images. The internal noise component was incorporated to model the inefficiency and variability of human observer (HO) performance, and to generate the ultimate DL-MO test statistics. Seven abdominal CT exams were retrospectively collected from the same type of CT scanners. To compare DL-MO with HO, 12 experimental conditions with varying lesion size, lesion contrast, radiation dose, and reconstruction types were generated, each condition with 154 trials. CT images of a real liver metastatic lesion were numerically modified to generate lesion models with four lesion sizes (5, 7, 9, and 11mm) and three contrast levels (15, 20, and 25HU). The lesions were inserted into patient liver images using a projection-based method. A validated noise insertion tool was used to synthesize CT exams with 50% and 25% of routine radiation dose level. CT images were reconstructed using the weighted filtered back projection algorithm and an iterative reconstruction algorithm. Four medical physicists performed a two-alternative forced choice (2AFC) detection task (with multislice scrolling viewing mode) on patient images across the 12 experimental conditions. DL-MO was operated on the same datasets. Statistical analyses were performed to evaluate the correlation and agreement between DL-MO and HO. A statistically significant positive correlation was observed between DL-MO and HO for the 2AFC low-contrast detection task that involves patient liver background. The corresponding Pearson product moment correlation coefficient was 0.986 [95% confidence interval (0.950, 0.996)]. Bland-Altman agreement analysis did not indicate statistically significant differences. The proposed DL-MO is highly correlated with HO in a low-contrast detection task that involves realistic patient liver background. This study demonstrated the potential of the proposed DL-MO to assess image quality directly based on patient images in realistic, clinically relevant CT tasks.

1H NMR metabolomic analysis of skin and blubber of bottlenose dolphins reveals a functional metabolic dichotomy

The common bottlenose dolphin (Tursiops truncatus) is a carnivorous cetacean that thrives in marine environments, one of the apex predators of the marine food web. They are found in coastal and estuarine ecosystems, which are known to be sensitive to environmental impacts. Dolphins are considered sentinel organisms for monitoring the health of coastal marine ecosystems due to their role as predators that can bioaccumulate contaminants. Although recent studies have focused on capturing the circulating metabolomes of these mammals, and in the context of pollutants and exposures in the marine environment, skin and blubber are important surface and protective tissues that have not been adequately probed for metabolism. Using a proton nuclear magnetic resonance spectroscopy (1H NMR) based metabolomics approach, we quantified 51 metabolites belonging to 74 different metabolic pathways in the skin and blubber of stranded bottlenose dolphin (n = 4) samples collected at different localities in the Southern Zone coast of Yucatan Peninsula of Mexico. Results indicate that metabolism of skin and blubber are quantitatively very different. These metabolite abundances could help discriminate the tissue-types using supervised partial least square regression discriminant analysis (PLSDA). Further, using hierarchical clustering analysis and random forest analysis of the metabolite abundances, the results pointed to unique metabolites that are important classifiers of the tissue-type. On one hand, the differential metabolic patterns, mainly linking fatty acid metabolism and ketogenic amino acids, seem to constitute a characteristic of blubber, thus pointing to fat synthesis and deposition. On the other hand, the skin showed several metabolites involved in gluconeogenic pathways, pointing towards an active anabolic energy-generating metabolism. The most notable pathways found in both tissues included: urea cycle, nucleotide metabolism, amino acid metabolism, glutathione metabolism among others. Our 1H NMR metabolomics analysis allowed the quantification of metabolites associated with these two organs, i.e., pyruvic acid, arginine, ornithine, 2-hydroxybutyric acid, 3-hydroxyisobutyric acid, and acetic acid, as discriminatory and classifying metabolites. These results would lead to further understanding of the functional and physiological roles of dolphin skin and blubber metabolism for better efforts in their conservation, as well as useful target biopsy tissues for monitoring of dolphin health conditions in marine pollution and ecotoxicology studies.

Early detection of virus infection in live human cells using Raman spectroscopy.

Virus infection of a human cell was determined only 3h after invagination. We used viral vector Ad-CMV-control (AdC), which lacks the E1 gene coding for early polypeptide 1 (E1). AdC can replicate in human embryonic kidney 293 (HEK293) cells into which the E1 gene has been transfected. According to partial least-square regression discriminant analysis, it was assumed that two kinds of reaction take place in the cell during viral invasion. The first response of the cell was determined 3h after the virus invasion, and the second one was determined ∼9 h later. The first one seems to be due to compositional changes in DNA. Analysis of large-scale datasets strongly indicated that the second reaction can be attributed to a reduction in protein concentration or uptake of phenylalanine into the nucleus.

Erratum to: Rapid discrimination of F1 hybrid seeds from their parental lines and selection of protein-rich corn lines for silage corn breeding using FT-IR spectroscopy combined by multivariate analysis

This study aims to establish the discrimination of F1 hybrid seeds from their parental lines and rapid selection of proteinrich lines from inbreeding lines of corn using FT-IR spectroscopy combined by multivariate analysis. Eight individual seeds from the maternal and paternal lines of Gwangpyeongok and their F1 progeny seeds were subjected to FT-IR spectroscopy. A total of 176 corn inbreeding lines including commercial corn cultivars were subjected to FT-IR spectroscopy. To establish the prediction model for total protein content from corn seed, 33 corn inbreeding lines out of 176 were randomly selected and total seed protein contents using Bradford assay were examined. PLS-DA (partial least square regression discriminant analysis) could clearly discriminate F1 hybrid seeds from their parental lines. PC (principal component) loading values show that 1,700 – 1,500 cm−1 and 1,200 – 900 cm−1 regions of FT-IR spectra are significantly important for discrimination of corn lines. The prediction model for total protein contents was established by PLS (partial least square regression) algorithm, and its accuracy was confirmed by cross-validation test (R2 = 0.94). After external validation fromexternal 25 corn inbreeding lines, regression coefficient (R2) was 0.78 which indicated that the prediction model had relatively good accuracy. Thus, considering these results we suggest that FT-IR combined with multivariate analysis could be applied as a novel tool for high-throughput screening of F1 hybrid seeds from their parental lines and protein-rich lines for breeding of silage corn cultivar.

Metabolic Profiling of the Mexican Anxiolytic and Sedative PlantGalphimia glaucaUsing Nuclear Magnetic Resonance Spectroscopy and Multivariate Data Analysis

Galphimia glauca is popularly employed in Mexico for the treatment of central nervous system disorders. Pharmacological and phytochemical studies have resulted in the identification of the anxiolytic and sedative principle consisting of a mixture of nor-secofriedelanes, named the galphimine series (1 - 9). These active constituents were found in plants collected in the vicinity of a restricted region in Central Mexico, where this species is abundant. A metabolic profiling carried out by means of 1H-NMR spectroscopy and multivariate data analysis was applied to crude extracts from wild plant populations, collected from six different locations as a quality control assessment, in order to differentiate their chemical profile. Principal component analysis (PCA) of the 1H-NMR spectra revealed clear variations among the populations, with two populations out of the six studied manifesting differences, when the principal components PC-1 and PC-2 were analyzed. These two PCs permitted the differentiation of the various sample populations, depending on the presence of galphimines. This information consistently correlated with the corresponding HPLC analysis. The neuropharmacological effects of the crude extracts were evaluated by using ICR mice in the elevated plus maze, as well as the sodium pentobarbital-induced hypnosis models. Both assays demonstrated anxiolytic and sedative responses only among those sample populations which had previously been differentiated by PC-1. Partial least square regression-discriminant analysis (PLS-DA) also confirmed a strong correlation between the observed effects and the metabolic profiles of the plants. The overall results of this study confirm the benefits of using metabolic profiling for the in silico analysis of active principles in medicinal plants.

Composition of the bacterial population in sand-filter columns receiving artificial wastewater, evaluated by soft independent modelling of class analogy (SIMCA)

The objective of the present study was to develop a test system for clustering and characterizing the functions of the bacterial populations in wastewater infiltration systems. To create different bacterial ecosystems, sand-filter columns were loaded with a low loading (2.9 ± 0.1 cm d −1) and a high loading (8.7 ± 0.2 cm d −1) of an artificial wastewater for 114 days. Before loading onto the columns the wastewater was pretreated in a three-chambered container for 24 h. The low loaded filters attained a high and steady rate of mineralization and complete nitrification, whereas the high loaded filters developed a clogging surface, and incomplete nitrification despite a high rate of mineralization. On day 60 the sand surface was sampled, and the bacteria were extracted and enumdrated. From each column system 88 bacterial isolates were collected. Each isolate was subjected to 51 physiological and biochemical tests. The results were analyzed using the SIMCA statistical package. Using partial least-square regression discriminant analysis we developed a model that revealed five major clusters of bacteria among the 176 isolates. Two clusters consisted of bacteria from the high loaded filters, and another two clusters consisted of bacteria from the low loaded filters. The fifth cluster represented a mixed population containing members from both loading levels. The results were used to characterize the bacterial populations, and a hypothesis as to how these populations developed in their respective system was proposed. The model could be used for classifying the bacterial isolates into low or high loaded system. Only six isolates from each loading level were misclassified. The model was stable as indicated by the fact that both the numbers of tests and the numbers of isolates entered in the model could be reduced without any significant loss of information.