Guide Metabolome Analyses:: Strategies for Systems Biology

Free download. Book file PDF easily for everyone and every device. You can download and read online Metabolome Analyses:: Strategies for Systems Biology file PDF Book only if you are registered here. And also you can download or read online all Book PDF file that related with Metabolome Analyses:: Strategies for Systems Biology book. Happy reading Metabolome Analyses:: Strategies for Systems Biology Bookeveryone. Download file Free Book PDF Metabolome Analyses:: Strategies for Systems Biology at Complete PDF Library. This Book have some digital formats such us :paperbook, ebook, kindle, epub, fb2 and another formats. Here is The CompletePDF Book Library. It's free to register here to get Book file PDF Metabolome Analyses:: Strategies for Systems Biology Pocket Guide.
Metabolome Analyses:: Strategies for Systems Biology 1st Edition by Vaidyanathan, Seetharaman published by Springer Hardcover Hardcover – April 28,
Table of contents

As a word of caution, it is important to consider that the results of network inference and data analysis in general can be affected by data pre-treatment also known as pre-processing such as scaling, transformation and normalization.

Such pre-treatments are routinely applied to metabolomics data in order to correct for systematic and unwanted variation such as sample-to sample to variability induced by dilution effects e. The literature on the topic is huge: we refer the reader to Bijlsma et al. DiLeo et al.

Systems Biology and Multi-Omics Integration: Viewpoints from the Metabolomics Research Community

With this approach, they could recognize and model systems-level differences in biological networks even where the poorly defined phenotypes precluded the use of PCA or other multivariate approaches. Lusczek et al. They could define network modules i. Within those modules they identified hub metabolites related to cellular respiration, highlighting its fundamental role in the pathophysiology of haemorrhagic shock and to late resuscitation time points.

They observed that PLS discriminant analysis PLS-DA did not capture the significance of several hub metabolites, which emerged only in the network analysis.

Supporting Information

In the same work Lusczek et al. Such limitations rest on the assumptions that the network shows a scale-free topology, that is with few metabolites highly connected and many metabolites with low connectivity; this translates in the connectivity P k and the clustering coefficient C k to follow a power law.

The authors found P k to follow a power law but not C k , indicating the absence of modular structure in the network of urinary metabolites. A further hypothesis put forward in the same work was that networks constructed from metabolite profiles derived from biological samples that are metabolically active, such as blood or tissue, may exhibit power law i. However, in contrast to gene regulatory network, expression networks or metabolic networks, the metabolite correlation networks have not been fully characterized in terms of network topology i.

Therefore, it is not very clear what are the expected or more likely network properties e. We refer the reader to Lee et al. The weighted adjacency matrix is transformed into a binary topological matrix by additionally imposing a threshold on the mutual information.

The threshold is usually 0, leading to all non-zero values being transformed to 1. The weights constitute a probabilistic measurement of edge likeliness on which a threshold can be applied to obtain a binarized association network. This algorithm was used to construct association networks of blood metabolites characteristics of low and high latent cardiovascular risk Saccenti et al. They modelled the subject-specific networks through a statistical mechanics approach Menichetti et al.

The same approach was used in a study aiming to compare metabolite association networks obtained from serum and plasma samples. The networks were found to be topologically similar but showed local differences as in the case of amino acids see Fig.

Similarly, Vignoli et al. In particular, they investigated the different patterns of interconnectedness and observed sex-related variability in several metabolic pathways branched-chain amino acids, ketone bodies and propanoate metabolism as well as reduction in the connectivity of metabolite hubs linked to age in both sex groups.

Each dot represents a network that corresponds to a given cardiovascular CVD risk parameter. The associated CVD risk parameters are indicated in upper case for high risk and lower case for low risk. Reproduced with permission from Saccenti et al. Pirhaij et al. Grounding on database information, the algorithm infers the identity of unassigned metabolites corresponding to features and the molecular mechanisms underlying their dysregulation. This innovative approach helps to reduce the bias towards well-studied metabolites typical of targeted metabolomics.

The algorithm takes as input a list of LC—MS peaks that differ between two different conditions and searches for them in a databases containing over 42, nodes either proteins or metabolites connected by over one million weighted edges representing interactions between proteins as well as enzymatic and transporter reactions. The metabolism is a network structure that can be approached as a system of interdependent variables that enable mathematical modelling through kinetic models.

These models are defined as systems of ordinary differential equations describing the time course of metabolite concentrations as a function of rate laws that account for enzyme catalysis. The development of these models requires to know both the network structure and the reaction kinetics and parameters Klipp et al. On the one hand, there is a large accumulated knowledge regarding the network structure, which is stored in databases like KEGG Kanehisa et al.

Although this is a well-studied cellular level, the true structures can be importantly affected by factors like compartmentalization de Mas et al. On the other hand, regarding reaction kinetics, there is also an accumulated knowledge, which can be explored in databases such as BRENDA Scheer et al. In addition, the available measurements of the kinetic properties of enzymes historically come from systems reconstituted in vitro using purified enzymes Savageau In this setting, the ideal conditions of homogeneity and free diffusion are fulfilled, and consequently the resulting models may neglect some factors affecting the kinetic properties, such as molecular crowding Schnell and Turner and limited diffusion Alekseev et al.

To overcome these limitations, alternative approaches combine sampling methods with the integration of systemic available data and in vivo observations fluxes, concentrations, perturbation experiments, … Andreozzi et al. Alternative approaches take advantage of the current availability of data regarding the network structure and of the lineal nature of the system used to describe it, to apply optimization techniques to infer flux distributions Fouladiha and Marashi Genome—scale models accounting for thousands of reactions are currently available Chelliah et al. For those models including only the network structure as well as for complete kinetic models, it is useful to adopt techniques based on stable isotopes to know about the internal distribution of the metabolism.

These are addressed in the next section. Although the analysis of metabolite correlative networks may not grasp the complete underlying metabolic mechanisms, it is certainly a valuable tool for the exploration of metabolomics data, as shown by the budding literature on the topic. The use of stable isotopes can provide a greater insight on the mechanisms that underlie the observed metabolomics profiles, permitting a direct analysis of mechanistic changes in metabolism. Each chemical reaction or transport process involved in a metabolic pathway is associated with a rate flux of transformation or transport.

Mechanistic changes at the level of the metabolism are likely to produce changes in the distribution of fluxes. Intracellular fluxes are not directly measurable, but the use of stable isotope-enriched nutrients, such as 1,2- 13 C 2 -glucose or 13 C 5 , 15 N 2 -glutamine, in in cell culture media and the application of Stable Isotope Resolved Metabolomics SIRM Fan et al.

This can be used to estimate information about fluxes, such as their relative or absolute magnitudes Lee ; Zamboni et al. The estimation of fluxes based on the measured patterns of stable isotope labeling especially using 13 C relies upon a combination of different methods, going from the direct interpretation of the labeling patterns to computational model-based approaches Buescher et al. Frequently, direct interpretation of labeling patterns is sufficient to provide information on the relative activities of pathways, on qualitative changes in pathway contributions via alternative metabolic routes, and on nutrient contribution to the production of different metabolites Buescher et al.

A recent example is the direct interpretation of the contributions of isotopic labeling tracers like 1,2- 13 C 2 -glucose to the synthesys of pentoses phosphate Dong et al. During the last years, the use of this and other isotopic labelling tracers have been applied to unveil the different metabolic pathways activated in cancer cells see for a review Dong et al. By using computational approaches, all internal metabolic fluxes can be estimated simultaneously by combining the measured labeling patterns resulting from isotope propagation with the measured cellular uptake and secretion rates Buescher et al.


  1. A Promise Made (Family Feud, Book 2)?
  2. Log in to Wiley Online Library.
  3. Introduction.
  4. Strategies for Systems Biology!

A reliable model of the relevant network of biochemical reactions is an indispensable input to the computational approach. The reliability of hypotheses regarding flux distributions can be evaluated by comparing measured and predicted isotopologue distributions. In many cases, a system of balance equations around isotopomers—which depend on specific fluxes—is solved to predict label enrichments. Fluxes are iteratively changed until the difference among measured and predicted label enrichments is reduced. Overview of metabolic flux modelling using stable isotope resolved metabolomics data.

Ideally, assuming steady state, the distribution of isotopologues would only depend on the distribution of fluxes and the labeled and non-labeled status of the substrates used in the experiment. However, 13 C propagation from tracer precursors to products is a dynamic phenomenon. Progressively, these products are enriched in 13 C, with concomitant decrease in M0. Isotopic steady state Selivanov et al.

For these larger pools, M0 values are oversized and may not decrease to the hypothetical value that should be reached at steady state. Accordingly, as an alternative, some software platforms allow for solving the fitting procedure under non isotopic steady state e. Isodyn, INCA among those cited above. Enrichment analysis as applied in metabolomics is largely based on the approaches implemented for the analysis of transcriptomes, known as Gene Set Enrichment Analysis GSEA Subramanian et al.

Recommended for you

In practice, the goal of the approach is to detect biological processes, such as metabolic pathways, that differ in the experimental dataset of interest versus control datasets. Replacing gene transcription level with alterations in metabolite concentrations provides a very straightforward approach to interpret metabolomics experiments in terms of changes in the activity of cellular processes. Biochemical genetic and genomic knowledgebase of large scale metabolic reconstructions.

Schellenberger et al. A collection of computationally predicted metabolic pathways for nearly organisms whose genome is available. A partially curated database of metabolic reactions derived from the human genome. A curated, peer-reviewed knowledgebase of biological pathways, including metabolic pathways. It is mainly focused on human pathways. A database of biological pathways maintained by and for the scientific community. A related approach is the so called over-representation analysis ORA, sometimes called annotation enrichment analysis where one checks whether a group of differentially expressed genes is enriched for a pathway or ontology term by using overlap statistics such as the cumulative hypergeometric distribution Doniger et al.

Bonferroni , allows researchers to evaluate whether specific pathways containing metabolites in an experiment-derived list are overrepresented. If the input list contains metabolites featuring different concentrations in different phenotypes e. As the first step of the analysis, metabolites are assigned to specific sets based on one or more reference databases. A group of metabolites are assigned to the same set if they are known to be: i involved in the same biological processes i.

Different strategies exist for performing MSEA depending, among others, on the statistical test applied. In the popular Globaltest method Goeman et al. The question whether these metabolites behave differently in the two conditions being compared can be translated into the question whether the metabolite levels are predictive for the outcome Fig. The null hypothesis tested is that no metabolite in the pathway has a different concentration in the two conditions. Thus, the regression coefficients are all zero if the group of selected metabolites has no influence on the phenotype.

Unfortunately, the number of coefficients is often much larger than the number of samples leaving no room for classical testing procedures. Goeman et al. Thus, the test becomes whether the covariance is zero null hypothesis or different from zero alternative hypothesis. A correction is needed for multiple hypothesis pathway testing e.

Metabolome Analyses:

The Globaltest detects consistent differences in patterns of metabolite levels between two conditions. It does not test in which direction a pathway is regulated up or down , nor it determines how many metabolites have changed concentration levels between two conditions.

If the tested pathway is activated or inhibited by the tested condition e. However, the results may change, depending on which metabolites are included, i.