Robert A. van den Berg

Centering, scaling, and transformations: improving the biological information content of metabolomics data.

BackgroundExtracting relevant biological information from large data sets is a major challenge in functional genomics research. Different aspects of the data hamper their biological interpretation. For instance, 5000-fold differences in concentration for different metabolites are present in a metabolomics data set, while these differences are not proportional to the biological relevance of these metabolites. However, data analysis methods are not able to make this distinction. Data pretreatment methods can correct for aspects that hinder the biological interpretation of metabolomics data sets by emphasizing the biological information in the data set and thus improving their biological interpretability.ResultsDifferent data pretreatment methods, i.e. centering, autoscaling, pareto scaling, range scaling, vast scaling, log transformation, and power transformation, were tested on a real-life metabolomics data set. They were found to greatly affect the outcome of the data analysis and thus the rank of the, from a biological point of view, most important metabolites. Furthermore, the stability of the rank, the influence of technical errors on data analysis, and the preference of data analysis methods for selecting highly abundant metabolites were affected by the data pretreatment method used prior to data analysis.ConclusionDifferent pretreatment methods emphasize different aspects of the data and each pretreatment method has its own merits and drawbacks. The choice for a pretreatment method depends on the biological question to be answered, the properties of the data set and the data analysis method selected. For the explorative analysis of the validation data set used in this study, autoscaling and range scaling performed better than the other pretreatment methods. That is, range scaling and autoscaling were able to remove the dependence of the rank of the metabolites on the average concentration and the magnitude of the fold changes and showed biologically sensible results after PCA (principal component analysis).In conclusion, selecting a proper data pretreatment method is an essential step in the analysis of metabolomics data and greatly affects the metabolites that are identified to be the most important.

Systems analysis of protective immune responses to RTS,S malaria vaccination in humans

Significance The RTS,S malaria vaccine is the most advanced malaria vaccine candidate to be tested in humans. Despite its promise, there is little understanding of its mechanism of action. In this work, we describe the use of a systems biological approach to identify “molecular signatures” that are induced rapidly after the standard RTS,S vaccination regimen, consisting of three RTS,S immunizations, or with a different regimen consisting of a primary immunization with recombinant adenovirus 35 (Ad35) expressing the circumsporozoite malaria antigen followed by two immunizations with RTS,S. These results reveal important insights about the innate and adaptive responses to vaccination and identify signatures of protective immunity against malaria. RTS,S is an advanced malaria vaccine candidate and confers significant protection against Plasmodium falciparum infection in humans. Little is known about the molecular mechanisms driving vaccine immunity. Here, we applied a systems biology approach to study immune responses in subjects receiving three consecutive immunizations with RTS,S (RRR), or in those receiving two immunizations of RTS,S/AS01 following a primary immunization with adenovirus 35 (Ad35) (ARR) vector expressing circumsporozoite protein. Subsequent controlled human malaria challenge (CHMI) of the vaccinees with Plasmodium-infected mosquitoes, 3 wk after the final immunization, resulted in ∼50% protection in both groups of vaccinees. Circumsporozoite protein (CSP)-specific antibody titers, prechallenge, were associated with protection in the RRR group. In contrast, ARR-induced lower antibody responses, and protection was associated with polyfunctional CD4+ T-cell responses 2 wk after priming with Ad35. Molecular signatures of B and plasma cells detected in PBMCs were highly correlated with antibody titers prechallenge and protection in the RRR cohort. In contrast, early signatures of innate immunity and dendritic cell activation were highly associated with protection in the ARR cohort. For both vaccine regimens, natural killer (NK) cell signatures negatively correlated with and predicted protection. These results suggest that protective immunity against P. falciparum can be achieved via multiple mechanisms and highlight the utility of systems approaches in defining molecular correlates of protection to vaccination.

A flexible framework for sparse simultaneous component based data integration

Abstract1 BackgroundHigh throughput data are complex and methods that reveal structure underlying the data are most useful. Principal component analysis, frequently implemented as a singular value decomposition, is a popular technique in this respect. Nowadays often the challenge is to reveal structure in several sources of information (e.g., transcriptomics, proteomics) that are available for the same biological entities under study. Simultaneous component methods are most promising in this respect. However, the interpretation of the principal and simultaneous components is often daunting because contributions of each of the biomolecules (transcripts, proteins) have to be taken into account.2 ResultsWe propose a sparse simultaneous component method that makes many of the parameters redundant by shrinking them to zero. It includes principal component analysis, sparse principal component analysis, and ordinary simultaneous component analysis as special cases. Several penalties can be tuned that account in different ways for the block structure present in the integrated data. This yields known sparse approaches as the lasso, the ridge penalty, the elastic net, the group lasso, sparse group lasso, and elitist lasso. In addition, the algorithmic results can be easily transposed to the context of regression. Metabolomics data obtained with two measurement platforms for the same set of Escherichia coli samples are used to illustrate the proposed methodology and the properties of different penalties with respect to sparseness across and within data blocks.3 ConclusionSparse simultaneous component analysis is a useful method for data integration: First, simultaneous analyses of multiple blocks offer advantages over sequential and separate analyses and second, interpretation of the results is highly facilitated by their sparseness. The approach offered is flexible and allows to take the block structure in different ways into account. As such, structures can be found that are exclusively tied to one data platform (group lasso approach) as well as structures that involve all data platforms (Elitist lasso approach).4 AvailabilityThe additional file contains a MATLAB implementation of the sparse simultaneous component method.

Metabolomics data exploration guided by prior knowledge

In metabolomics research, it is often important to focus the data analysis to specific areas of interest within the metabolome. In this paper, we describe the application of consensus principal component analysis (CPCA) and canonical correlation analysis (CCA) as a means to explore the relation between metabolome data and (i) biochemically related metabolites and (ii) an amino acid biosynthesis pathway. CPCA searches for major trends in the behavior of metabolite concentrations that are in common for the metabolites of interest and the remainder of the metabolome. CCA identifies the strongest correlations between the metabolites of interest and the remainder of the metabolome. CPCA and CCA were applied to two different microbial metabolomics data sets. The first data set, derived from Pseudomonas putida S12, was relatively simple as it contained metabolomes obtained under four environmental conditions only. The second data set, obtained from Escherichia coli, was much more complex as it consisted of metabolomes obtained under 28 different environmental conditions. In case of the simple and coherent P. putida S12 data set, CCA and CPCA gave similar results as the variation in the subset of the selected metabolites and the remainder of the metabolome was similar. In contrast, CCA and CPCA yielded different results in case of the E. coli data set. With CPCA the trends in the selected subset--the phenylalanine biosynthesis pathway--dominated the results. The main trends were related to high and low phenylalanine productivity, and the metabolites showing a similar behavior in concentration were metabolites regulating the phenylalanine biosynthesis route in the subset and metabolites related to general amino acid metabolism in the remainder of the metabolome. With CCA, neither subset truly dominated the data analysis. CCA described the differences between the wild type and the overproducing strain and the differences between the succinate and glucose grown cells. For the difference between the wild type and the overproducing strain, metabolites from the beginning and the end of aromatic amino acid pathways like erythrose-4-phosphate, tryptophan, and phenylalanine were important for the selected metabolites. CCA and CPCA proved to be complementary data analysis tools that enable the focusing of the data analysis on groups of metabolites that are of specific interest in relation to the remainder of the metabolome. Compared to an ordinary PCA, focusing the data analysis on biologically relevant metabolites lead especially for the complex E. coli data to a better biological interpretation of the data.

Simultaneous analysis of coupled data matrices subject to different amounts of noise

In many areas of science, research questions imply the analysis of a set of coupled data blocks, with, for instance, each block being an experimental unit by variable matrix, and the variables being the same in all matrices. To obtain an overall picture of the mechanisms that play a role in the different data matrices, the information in these matrices needs to be integrated. This may be achieved by applying a data-analytic strategy in which a global model is fitted to all data matrices simultaneously, as in some forms of simultaneous component analysis (SCA). Since such a strategy implies that all data entries, regardless the matrix they belong to, contribute equally to the analysis, it may obfuscate the overall picture of the mechanisms underlying the data when the different data matrices are subject to different amounts of noise. One way out is to downweight entries from noisy data matrices in favour of entries from less noisy matrices. Information regarding the amount of noise that is present in each matrix, however, is, in most cases, not available. To deal with these problems, in this paper a novel maximum-likelihood-based simultaneous component analysis method, referred to as MxLSCA, is proposed. Being a stochastic extension of SCA, in MxLSCA the amount of noise in each data matrix is estimated and entries from noisy data matrices are downweighted. Both in an extensive simulation study and in an application to data stemming from cross-cultural emotion psychology, it is shown that the novel MxLSCA strategy outperforms the SCA strategy with respect to disclosing the mechanisms underlying the coupled data.

Illness and symptom perception: a theoretical approach towards an integrative measurement model.

Several models have been proposed to conceptualize psychological representations of health, illness, and bodily sensations. These models differ as to the cognitive and affective components they include, whether they study the interaction of these components, and whether associations between psychological representations of bodily states and affective and behavioral reactions to these representations are considered conditional. These different conceptualizations and corresponding measurement approaches exist in parallel without resulting in synergistic effects or theoretical advancements within the field. In this paper, we review theoretical models on perception and attitudes and construct an integrative theoretical framework on psychological representation of bodily symptoms as well as more abstract representations of health and disease. The aim of this combination of approaches is to unify the strengths of different research domains in the conceptualization and measurement of mental representations of bodily states. Furthermore, the aim is to specify new, testable predictions and implications about the (conditional) relationship of these mental representations and affective and behavioral consequences. A core element in this integrative model is comparison. We review how comparison processes can change the cognitive and affective reference frame for illness and symptom perception and in turn affective and behavioral reactions. We discuss implications for measurement of illness and symptom representations as well as implications for clinical practice. Finally, we make suggestions for a research agenda to validate the proposed model as well as to address new questions derived from it.

Cellular and molecular synergy in AS01-adjuvanted vaccines results in an early IFNγ response promoting vaccine immunogenicity.

Combining immunostimulants in adjuvants can improve the quality of the immune response to vaccines. Here, we report a unique mechanism of molecular and cellular synergy between a TLR4 ligand, 3-O-desacyl-4’-monophosphoryl lipid A (MPL), and a saponin, QS-21, the constituents of the Adjuvant System AS01. AS01 is part of the malaria and herpes zoster vaccine candidates that have demonstrated efficacy in phase III studies. Hours after injection of AS01-adjuvanted vaccine, resident cells, such as NK cells and CD8+ T cells, release IFNγ in the lymph node draining the injection site. This effect results from MPL and QS-21 synergy and is controlled by macrophages, IL-12 and IL-18. Depletion strategies showed that this early IFNγ production was essential for the activation of dendritic cells and the development of Th1 immunity by AS01-adjuvanted vaccine. A similar activation was observed in the lymph node of AS01-injected macaques as well as in the blood of individuals receiving the malaria RTS,S vaccine. This mechanism, previously described for infections, illustrates how adjuvants trigger naturally occurring pathways to improve the efficacy of vaccines.Adjuvants: Vaccine components working in synergy to improve beneficial effects of vaccinationA mechanism is revealed by which vaccine components co-operate to stimulate the immune system and improve vaccine efficacy. Some vaccines are formulated with adjuvants—compounds that induce a greater immune response to the vaccine and help to elicit greater protection against future infections. Arnaud Didierlaurent and his team of researchers at GSK Vaccines, Belgium, demonstrate that the two immunostimulants in the adjuvant AS01, used in several recently developed vaccines, works in tandem to trigger the activation of important immune system moderators. The synergistic effect of the immunostimulants modulate specific immune cells at the site of the vaccination to better prepare the body against future infection. Studies such as this allow us to better understand how vaccines work and lay the foundation for more informed research into future vaccine development.

Integrating functional genomics data using maximum likelihood based simultaneous component analysis

BackgroundIn contemporary biology, complex biological processes are increasingly studied by collecting and analyzing measurements of the same entities that are collected with different analytical platforms. Such data comprise a number of data blocks that are coupled via a common mode. The goal of collecting this type of data is to discover biological mechanisms that underlie the behavior of the variables in the different data blocks. The simultaneous component analysis (SCA) family of data analysis methods is suited for this task. However, a SCA may be hampered by the data blocks being subjected to different amounts of measurement error, or noise. To unveil the true mechanisms underlying the data, it could be fruitful to take noise heterogeneity into consideration in the data analysis. Maximum likelihood based SCA (MxLSCA-P) was developed for this purpose. In a previous simulation study it outperformed normal SCA-P. This previous study, however, did not mimic in many respects typical functional genomics data sets, such as, data blocks coupled via the experimental mode, more variables than experimental units, and medium to high correlations between variables. Here, we present a new simulation study in which the usefulness of MxLSCA-P compared to ordinary SCA-P is evaluated within a typical functional genomics setting. Subsequently, the performance of the two methods is evaluated by analysis of a real life Escherichia coli metabolomics data set.ResultsIn the simulation study, MxLSCA-P outperforms SCA-P in terms of recovery of the true underlying scores of the common mode and of the true values underlying the data entries. MxLSCA-P further performed especially better when the simulated data blocks were subject to different noise levels. In the analysis of an E. coli metabolomics data set, MxLSCA-P provided a slightly better and more consistent interpretation.ConclusionMxLSCA-P is a promising addition to the SCA family. The analysis of coupled functional genomics data blocks could benefit from its ability to take different noise levels per data block into consideration and improve the recovery of the true patterns underlying the data. Moreover, the maximum likelihood based approach underlying MxLSCA-P could be extended to custom-made solutions to specific problems encountered.

Predicting RTS,S vaccine-mediated protection from transcriptomes in a malaria-challenge clinical trial

The RTS,S candidate malaria vaccine can protect against controlled human malaria infection (CHMI), but how protection is achieved remains unclear. Here, we have analyzed longitudinal peripheral blood transcriptome and immunogenicity data from a clinical efficacy trial in which healthy adults received three RTS,S doses 4 weeks apart followed by CHMI 2 weeks later. Multiway partial least squares discriminant analysis (N-PLS-DA) of transcriptome data identified 110 genes that could be used in predictive models of protection. Among the 110 genes, 42 had known immune-related functions, including 29 that were related to the NF-κB-signaling pathway and 14 to the IFN-γ-signaling pathway. Post-dose 3 serum IFN-γ concentrations were also correlated with protection; and N-PLS-DA of IFN-γ-signaling pathway transcriptome data selected almost all (44/45) of the representative genes for predictive models of protection. Hence, the identification of the NF-κB and IFN-γ pathways provides further insight into how vaccine-mediated protection may be achieved.

Different Adjuvants Induce Common Innate Pathways That Are Associated with Enhanced Adaptive Responses against a Model Antigen in Humans

To elucidate the role of innate responses in vaccine immunogenicity, we compared early responses to hepatitis B virus (HBV) surface antigen (HBsAg) combined with different Adjuvant Systems (AS) in healthy HBV-naïve adults, and included these parameters in multi-parametric models of adaptive responses. A total of 291 participants aged 18–45 years were randomized 1:1:1:1:1 to receive HBsAg with AS01B, AS01E, AS03, AS04, or Alum/Al(OH)3 at days 0 and 30 (ClinicalTrials.gov: NCT00805389). Blood protein, cellular, and mRNA innate responses were assessed at early time-points and up to 7 days after vaccination, and used with reactogenicity symptoms in linear regression analyses evaluating their correlation with HBs-specific CD4+ T-cell and antibody responses at day 44. All AS induced transient innate responses, including interleukin (IL)-6 and C-reactive protein (CRP), mostly peaking at 24 h post-vaccination and subsiding to baseline within 1–3 days. After the second but not the first injection, median interferon (IFN)-γ levels were increased in the AS01B group, and IFN-γ-inducible protein-10 levels and IFN-inducible genes upregulated in the AS01 and AS03 groups. No distinct marker or signature was specific to one particular AS. Innate profiles were comparable between AS01B, AS01E, and AS03 groups, and between AS04 and Alum groups. AS group rankings within adaptive and innate response levels and reactogenicity prevalence were similar (AS01B ≥ AS01E > AS03 > AS04 > Alum), suggesting an association between magnitudes of inflammatory and vaccine responses. Modeling revealed associations between adaptive responses and specific traits of the innate response post-dose 2 (activation of the IFN-signaling pathway, CRP and IL-6 responses). In conclusion, the ability of AS01 and AS03 to enhance adaptive responses to co-administered HBsAg is likely linked to their capacity to activate innate immunity, particularly the IFN-signaling pathway.