Ewoud J. J. van Velzen

Assessment of PLSDA cross validation

Classifying groups of individuals based on their metabolic profile is one of the main topics in metabolomics research. Due to the low number of individuals compared to the large number of variables, this is not an easy task. PLSDA is one of the data analysis methods used for the classification. Unfortunately this method eagerly overfits the data and rigorous validation is necessary. The validation however is far from straightforward. Is this paper we will discuss a strategy based on cross model validation and permutation testing to validate the classification models. It is also shown that too optimistic results are obtained when the validation is not done properly. Furthermore, we advocate against the use of PLSDA score plots for inference of class differences.

Metabolic fate of polyphenols in the human superorganism

Dietary polyphenols are components of many foods such as tea, fruit, and vegetables and are associated with several beneficial health effects although, so far, largely based on epidemiological studies. The intact forms of complex dietary polyphenols have limited bioavailability, with low circulating levels in plasma. A major part of the polyphenols persists in the colon, where the resident microbiota produce metabolites that can undergo further metabolism upon entering systemic circulation. Unraveling the complex metabolic fate of polyphenols in this human superorganism requires joint deployment of in vitro and humanized mouse models and human intervention trials. Within these systems, the variation in diversity and functionality of the colonic microbiota can increasingly be captured by rapidly developing microbiomics and metabolomics technologies. Furthermore, metabolomics is coming to grips with the large biological variation superimposed on relatively subtle effects of dietary interventions. In particular when metabolomics is deployed in conjunction with a longitudinal study design, quantitative nutrikinetic signatures can be obtained. These signatures can be used to define nutritional phenotypes with different kinetic characteristics for the bioconversion capacity for polyphenols. Bottom-up as well as top-down approaches need to be pursued to link gut microbial diversity to functionality in nutritional phenotypes and, ultimately, to bioactivity of polyphenols. This approach will pave the way for personalization of nutrition based on gut microbial functionality of individuals or populations.

Multivariate paired data analysis: multilevel PLSDA versus OPLSDA

Metabolomics data obtained from (human) nutritional intervention studies can have a rather complex structure that depends on the underlying experimental design. In this paper we discuss the complex structure in data caused by a cross-over designed experiment. In such a design, each subject in the study population acts as his or her own control and makes the data paired. For a single univariate response a paired t-test or repeated measures ANOVA can be used to test the differences between the paired observations. The same principle holds for multivariate data. In the current paper we compare a method that exploits the paired data structure in cross-over multivariate data (multilevel PLSDA) with a method that is often used by default but that ignores the paired structure (OPLSDA). The results from both methods have been evaluated in a small simulated example as well as in a genuine data set from a cross-over designed nutritional metabolomics study. It is shown that exploiting the paired data structure underlying the cross-over design considerably improves the power and the interpretability of the multivariate solution. Furthermore, the multilevel approach provides complementary information about (I) the diversity and abundance of the treatment effects within the different (subsets of) subjects across the study population, and (II) the intrinsic differences between these study subjects.

The metabolic fate of red wine and grape juice polyphenols in humans assessed by metabolomics.

The metabolic impact of polyphenol-rich red wine and grape juice consumption in humans was studied using a metabolomics approach. Fifty-eight men and women participated in a placebo-controlled, double-crossover study in which they consumed during a period of 4 wk, either a polyphenol-rich 2:1 dry mix of red wine and red grape juice extracts (MIX) or only a grape juice extract (GJX). Twenty-four-hour urine samples were collected after each intervention. (1)H NMR spectroscopy was applied for global metabolite profiling, while GC-MS was used for focused profiling of urinary phenolic acids. Urine metabolic profiles after intake of both polyphenol-rich extracts were significantly differentiated from placebo using multilevel partial least squares discriminant analysis. A significant 35% increase in hippuric acid excretion (p<0.001) in urine was measured after the MIX consumption as) or only a red grape juice dry extract (GJX). 24-h urine samples were collected after each intervention. 1H-NMR spectroscopy was applied for global metabolite profiling, while gas chromatography-mass spectrometry (GC-MS) was used for focused profiling of urinary phenolic acids. Urine metabolic profiles after intake of both polyphenol-rich extracts were significantly differentiated from placebo using multilevel partial least squares discriminant analysis (ML-PLS-DA). A significant 35% increase in hippuric acid excretion (p<0.001) in urine was measured after the MIX consumption compared with placebo, whereas no change was found after GJX consumption. GC-MS-based metabolomics of urine allowed identification of 18 different phenolic acids, which were significantly elevated following intake of either extract. Syringic acid, 3- and 4-hydroxyhippuric acid and 4-hydroxymandelic acid were the strongest urinary markers for both extracts. MIX and GJX consumption had a slightly different effect on the excreted phenolic acid profile and on endogenous metabolite excretion, possibly reflecting their different polyphenol composition.

Phenotyping Tea Consumers by Nutrikinetic Analysis of Polyphenolic End-Metabolites

An integration of metabolomics and pharmacokinetics (or nutrikinetics) is introduced as a concept to describe a human study population with different metabolic phenotypes following a nutritional intervention. The approach facilitates an unbiased analysis of the time-response of body fluid metabolites from crossover designed intervention trials without prior knowledge of the underlying metabolic pathways. The method is explained for the case of a human intervention study in which the nutrikinetic analysis of polyphenol-rich black tea consumption was performed in urine over a period of 48 h. First, multilevel PLS-DA analysis was applied to the urinary 1H NMR profiles to select the most differentiating biomarkers between the verum and placebo samples. Then, a one-compartment nutrikinetic model with first-order excretion, a lag time, and a baseline function was fitted to the time courses of these selected biomarkers. The nutrikinetic model used here fully exploits the crossover structure in the data by fitting the data from both the treatment period and the placebo period simultaneously. To demonstrate the procedure, a selected set of urinary biomarkers was used in the model fitting. These metabolites include hippuric acid, 4-hydroxyhippuric acid and 1,3-dihydroxyphenyl-2-O-sulfate and derived from microbial fermentation of polyphenols in the gut. Variations in urinary excretion between- and within the subjects were observed, and used to provide a phenotypic description of the test population.

GC-MS methods for metabolic profiling of microbial fermentation products of dietary polyphenols in human and in vitro intervention studies

Flavonoids, a subclass of polyphenols, are major constituents of many plant-based foods and beverages, including tea, wine and chocolate. Epidemiological studies have shown that a flavonoid-rich diet is associated with reduced risk of cardiovascular diseases. The majority of the flavonoids survive intact until they reach the colon where they are then extensively metabolized into smaller fragments. Here, we describe the development of GC-MS-based methods for the profiling of phenolic microbial fermentation products in urine, plasma, and fecal water. Furthermore, the methods are applicable for profiling products obtained from in vitro batch culture fermentation models. The methods incorporate enzymatic deconjugation, liquid-liquid extraction, derivatization, and subsequent analysis by GC-MS. At the level of individual compounds, the methods gave recoveries better than 80% with inter-day precision being better than 20%, depending on the matrix. Limits of detection were below 0.1 microg/ml for most phenolic acids. The newly developed methods were successfully applied to samples from human and in-vitro intervention trials, studying the metabolic impact of flavonoid intake. In conclusion, the methods presented are robust and generally applicable to diverse biological fluids. Its profiling character is useful to investigate on a large scale the gut microbiome-mediated bioavailability of flavonoids.

Discriminant Q2 (DQ2) for improved discrimination in PLSDA models

In this paper we introduce discriminant Q2 (DQ2) as an improvement for the Q2 value used in the validation of PLSDA models. DQ2 does not penalize class predictions beyond the class label value. With rigorous Monte Carlo simulations we show that when DQ2 is used, a smaller effect can be found statistically significant than when the standard Q2 is used.

Parameter selection for peak alignment in chromatographic sample profiling: objective quality indicators and use of control samples

In chromatographic profiling applications, peak alignment is often essential as most chromatographic systems exhibit small peak shifts over time. When using currently available alignment algorithms, there are several parameters that determine the outcome of the alignment process. Selecting the optimum set of parameters, however, is not straightforward, and the quality of an alignment result is at least partly determined by subjective decisions. Here, we demonstrate a new strategy to objectively determine the quality of an alignment result. This strategy makes use of a set of control samples that are analysed both spiked and non-spiked. With this set, not only the system and the method can be checked but also the quality of the peak alignment can be evaluated. The developed strategy was tested on a representative metabolomics data set using three software packages, namely Markerlynx™, MZmine and MetAlign. The results indicate that the method was able to assess and define the quality of an alignment process without any subjective interference of the analyst, making the method a valuable contribution to the data handling process of chromatography-based metabolomics data.

Factors Associated with Dough Stickiness as Sensed by Attenuated Total Reflectance Infrared Spectroscopy

ABSTRACT Attenuated total reflectance (ATR) and Fourier transform infrared (FTIR) spectroscopy have been applied in the characterization of sticky dough surfaces. The characterization provides insight in the chemical distribution of gluten protein, starch, water, and fat during dough kneading. ATR is especially useful for selective sampling of dough surfaces because the depth of penetration of radiation is quite shallow. For dough, it is calculated to be in the order of 0.5–4 μm in the mid-infrared, ideal for measurements of stickiness effects, where only the dough surface is of interest. To investigate the cohesive and adhesive properties of the individual dough constituents, dough was peeled from the ATR plate to study the material that adhered to it. The infrared spectra obtained indicate that fat and gluten protein appear to be located at the outer sticky dough surfaces, rather than water and starch. In comparison with gluten, the fatty component showed relatively strong adhesive forces to the ATR pla...

Cryptocalix[6]arenes; molecules with a large cavity

A new type of cavitand molecules with large cavities (4a–g) has been synthesized by the covalent three-point linking of a p-tert-butylcalix[6]arene to a cyclotriveratrylene (CTV) and their dynamic behavior has been studied by 1H NMR spectroscopy.