A Guideline to Univariate Statistical Analysis for LC/MS-Based Untargeted Metabolomics-Derived Data

Abstract

Several metabolomic software programs provide methods for peak picking, retention time alignment and quantification of metabolite features in LC/MS-based metabolomics. Statistical analysis, however, is needed in order to discover those features significantly altered between samples. By comparing the retention time and MS/MS data of a model compound to that from the altered feature of interest in the research sample, metabolites can be then unequivocally identified. This paper reports on a comprehensive overview of a workflow for statistical analysis to rank relevant metabolite features that will be selected for further MS/MS experiments. We focus on univariate data analysis applied in parallel on all detected features. Characteristics and challenges of this analysis are discussed and illustrated using four different real LC/MS untargeted metabolomic datasets. We demonstrate the influence of considering or violating mathematical assumptions on which univariate statistical test rely, using high-dimensional LC/MS datasets. Issues in data analysis such as determination of sample size, analytical variation, assumption of normality and homocedasticity, or correction for multiple testing are discussed and illustrated in the context of our four untargeted LC/MS working examples.

Nuclear magnetic resonance metabolomic profiling of urine provides a noninvasive alternative to the identification of biomarkers associated with endometriosis

“…However, a univariate statistical analysis of the data revealed that urine samples of women diagnosed with endometriosis exhibit differences in their metabolic profile compared with healthy women. The advantage of the univariate analysis over multivariate methods is the ease of application and results interpretation (26). The results showed that urine from endometriosis patients is characterized by increased concentrations of 4-Py, guanidinosuccinate, creatinine, taurine, valine, 2-hydroxyisovalerate, and U2 and decreased concentrations of lysine, U1, and U6.…”

Analysis of metabolomic data: tools, current strategies and future challenges for omics data integration

“…maximization of variance or correlation), and they do not take into account correlations from biological origin [17]. For this reason, the combined use of several statistical and data mining techniques is highly recommended [4]. Once identified, significantly expressed metabolites are first ranked using appropriate P-values; then, a cut-off threshold is applied to select the top-k ones from the ranked list.…”

“…The untargeted approach instead, also called 'metabolite fingerprinting', is not driven by an a priori hypothesis and it is used for complete metabolome comparison (i.e. as many metabolites as possible are measured and compared between samples) [4]. Metabolite variations are observed principally as total changes of chromatographic patterns without requiring previous knowledge of the compounds under investigation.…”

“…Although several literature about metabolomic data production and analysis techniques exist [3][4][5][15][16][17][18][19][20][21][22][23], to the best of our knowledge, a comprehensive review illustrating the available tools for the analysis and integration of metabolomic data with other 'omics' data has not been reported yet. Here, we describe and compare four tools (MetaCore TM , MetaboAnalyst, InCroMAP and 3Omics [24][25][26][27]), which we selected among the several ones currently available for metabolomic data analysis and integration (Table 1 have not yet overcome the previous ones.…”

Metabolomic Analysis of Schizosaccharomyces pombe: Sample Preparation, Detection, and Data Interpretation

“…Univariate and multivariate techniques are commonly applied to metabolome data sets (Liland 2011;Vinaixa et al 2012). An online service called MetaboAnalyst implements the common methods of statistical analysis and quality control, as well as metabolic pathway visualization (Xia et al 2012).…”