MetaboAnalyst 5.0 logo

MetaboAnalyst 5.0

Functional and statistical analysis of metabolomics data

Description

MetaboAnalyst 5.0 is a user-friendly web-based tool for statistical and functional interpretation of metabolomics (including lipidomics) data. The tool supports a wide-variety of univariate or multivariate statistical and machine learning methods to identify important features and patterns. MetaboAnalyst 5.0 also has methods for raw spectra processing, biomarker identification and meta-analysis. The tool supports Pathway Analysis for 26 organisms as well as Enrichment analysis of ~ 9, 000 metabolite sets, including all lipid classes from LipidMaps. Herein, the tool implements a smart-matching algorithm to aid users match their named-lipids with the internal MetaboAnalyst compound database. The tool provides rich and interactive visualizations to support users to explore their data. Finally, all tabular outputs, publication-quality figures, and PDF summary report are available at the end of a user’s session for download.

Technical Information

Download / Web-service link:
Yes
Source code repository:
NA
Programming languages:
R,
Java,
JavaScript
Platforms:
Windows,
Linux,
MacOS
Output formats:
CSV,
PNG,
PDF
Input formats:
CSV,
TXT,
mzML,
Peak list,
Compound list
Web platform:
Yes
Desktop client:
Yes
CLI:
No
GUI:
Yes
License:
GPL (Academic) / Commercial license (Commercial use)

Tasks

6) Analysis and visualization of lipidomics data
Direct connection to identification results:
Yes
Classification and feature selection:
Random Forest,
Support Vector Machine
Clustering and correlation analysis methods:
Hierarchical (dendrogram, heatmap),
Partitional (K-means, SOM) clustering
Feature identification:
Significance analysis of microarray (SAM), Empirical Bayesian Analysis of Microarray (EBAM)
Supervised multivariate statistical analysis methods:
PLS-DA,
sPLS-DA,
OrthoPLS-DA
Unsupervised multivariate statistical analysis methods:
PCA
Univariate statistical analysis methods:
T-test,
Fold-change analysis,
ANOVA,
Correlation heatmaps,
Pattern search,
Correlation networks
Missing data handling:
R-based methods
Data pre-treatments:
Filtering,
Normalization (Sample-specific, Sum, median, PQN, Pooled groups, Reference feature),
Scaling (mean centring, auto, pareto, range)