Dataset Information

Critical comparison of methods for fault diagnosis in metabolomics data.

ABSTRACT: Platforms like metabolomics provide an unprecedented view on the chemical versatility in biomedical samples. Many diseases reflect themselves as perturbations in specific metabolite combinations. Multivariate analyses are essential to detect such combinations and associate them to specific diseases. For this, usually targeted discriminations of samples associated to a specific disease from non-diseased control samples are used. Such targeted data interpretation may not respect the heterogeneity of metabolic responses, both between diseases and within diseases. Here we show that multivariate methods that find any set of perturbed metabolites in a single patient, may be employed in combination with data collected with a single metabolomics technology to simultaneously investigate a large array of diseases. Several such untargeted data analysis approaches have been already proposed in other fields to find both expected and unexpected perturbations, e.g. in Statistical Process Control. We have critically compared several of these approaches for their sensitivity and their correct identification of the specifically perturbed metabolites. Also a new approach is introduced for this purpose. The newly introduced Sparse Mean approach, which we find here as most sensitive and best able to identify the specifically perturbed metabolites, turns metabolomics into an untargeted diagnostic platform. Aside from metabolomics, the proposed approach may greatly benefit fault diagnosis with untargeted analyses in many other fields, such as Industrial Process Control, food Adulteration Detection, and Intrusion Detection.

SUBMITTER: Koeman M

PROVIDER: S-EPMC6362212 | biostudies-literature | 2019 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Critical comparison of methods for fault diagnosis in metabolomics data.

Koeman M M Engel J J Jansen J J Buydens L L

Scientific reports 20190204 1

Platforms like metabolomics provide an unprecedented view on the chemical versatility in biomedical samples. Many diseases reflect themselves as perturbations in specific metabolite combinations. Multivariate analyses are essential to detect such combinations and associate them to specific diseases. For this, usually targeted discriminations of samples associated to a specific disease from non-diseased control samples are used. Such targeted data interpretation may not respect the heterogeneity ...[more]

PMID: 30718783

Similar Datasets

Project description:Background: Integrated microbiome and metabolomics analyses hold the potential to reveal interactions between host and microbiota in relation to disease risks. However, there are few studies evaluating how field methods influence fecal microbiome characterization and metabolomics profiling. Methods: Five fecal collection methods [immediate freezing at -20°C without preservative, OMNIgene GUT, 95% ethanol, RNAlater, and Flinders Technology Associates (FTA) cards] were used to collect 40 fecal samples from eight healthy volunteers. We performed gut microbiota 16S rRNA sequencing, untargeted metabolomics profiling, and targeted metabolomics focusing on short chained fatty acids (SCFAs). Metrics included α-diversity and β-diversity as well as distributions of predominant phyla. To evaluate the concordance with the "gold standard" immediate freezing, the intraclass correlation coefficients (ICCs) for alternate fecal collection systems were calculated. Correlations between SCFAs and gut microbiota were also examined. Results: The FTA cards had the highest ICCs compared to the immediate freezing method for α-diversity indices (ICCs = 0.96, 0.96, 0.76 for Shannon index, Simpson's Index, Chao-1 Index, respectively), followed by OMNIgene GUT, RNAlater, and 95% ethanol. High ICCs (all >0.88) were observed for all methods for the β-diversity metric. For untargeted metabolomics, in comparison to immediate freezing which detected 621 metabolites at ≥75% detectability level, 95% ethanol showed the largest overlapping set of metabolites (n = 430; 69.2%), followed by FTA cards (n = 330; 53.1%) and OMNIgene GUT (n = 213; 34.3%). Both OMNIgene GUT (ICCs = 0.82, 0.93, 0.64) and FTA cards (ICCs = 0.87, 0.85, 0.54) had acceptable ICCs for the top three predominant SCFAs (butyric acid, propionic acid and acetic acid). Nominally significant correlations between bacterial genera and SCFAs (P < 0.05) were observed in fecal samples collected by different methods. Of note, a high correlation between the genus Blautia (known butyrate producer) and butyric acid was observed for both immediate freezing (r = 0.83) and FTA cards (r = 0.74). Conclusions: Four alternative fecal collection methods are generally comparable with immediate freezing, but there are differences in certain measures of the gut microbiome and fecal metabolome across methods. Choice of method depends on the research interests, simplicity of fecal collection procedures and ease of transportation to the lab, especially for large epidemiological studies.

Dataset Information

Critical comparison of methods for fault diagnosis in metabolomics data.

Publications

Critical comparison of methods for fault diagnosis in metabolomics data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets