MetaboAnalyst offers three types of enrichment analysis for targeted metabolomics, including
- Over Representation Analysis (ORA, input: significant compound names)
- Single Sample Profiling (SSP, input: a complete sample profile)
- Quantitative Enrichment Analysis (QEA, input: a complete concentration table)
QEA is based on compound concentration table. It is often claimed to be more sensitive than ORA and has the potential to discover “subtle but consistent” changes among compounds within the same biological pathway.
Many algorithms have been developed in the last decade for this type of enrichment analysis, the most famous being the Gene Set Enrichment Analysis. Many new and improved methods have since been implemented. When users upload concentration tables, GlobalTest is used for functional enrichment analysis; for pathway enrichment analysis, users can choose between GlobalTest or GlobalAncova.
Some important features about these two methods include that they support binary, multi-group, as well as continuous phenotypes, and p values can be approximated efficiently based on the asymptotic distribution without using permutations, which is critical for developing web applications.