Home

written by Don Teng, last updated 9 Jan 2020.

Statistics for "Semi-targeted" Analysis

"Semi-targeted" in the sense that a set of 200 - 300 reference metabolites are injected and assessed along with the samples, so as to get a more confident idea of compound identification/annotation further on, as opposed to a purely untargeted study which would only process the samples. Normalization is not included in this section; it has its own section below.

• Overview
• Basic Experimental Design
• Statistical power and sample size
• Quality checks - standard error of the mean (SEM), and coefficient of variation (CoV).
• Differential Abundance - t-tests and the Mann-Whitney U test.

Pathway Analysis

Useful Tools

Click through for detailed descriptions and tutorials.

• iPath3 - for an overall view of your metabolites of interest on a global metabolic map.
• MetaboAnalyst: pathway enrichment analysis - outputs a list of enriched pathways for a treatment group, relative to a control group.
• MetaboAnalyst: metabolite set enrichment analysis - outputs a list of enriched metabolite sets for a treatment group, relative to a control group.
• Pathview - multi-omics tool. Web app which maps genomic or metabolomic data onto a KEGG subnetwork. Not very reliable at the moment. The gage backend is quite versatile, but may be less applicable to metabolomics, in that gage is best used for datasets with tens of thousands of features (not unusual for RNA-seq or LFQ data); a few hundred features would be deemed insufficient.
• GAGE - Generally Applicable Gene-set Enrichment for Pathway Analysis. Slight misnomer (in terms of the term "gene-set" in the name); would work for any 'omics as long as certain criteria are fulfilled (thus the term "generally applicable"). Link to Bioconductor.
• SBGNview - Same functionality as pathview, but uses SBGN as a graphical display file format of choice. As of Feb 2020, SBGNview failed to install in a Mac (Catalina 10.15.3), Win10, or Ubuntu 18.04 LTS, so I'm just going to keep an eye on it.

Multi-omics Analysis

• mixOmics - (in progress) specifically the rCCA workflow.
• EGSEA - Enhanced gene set enrichment analysis. Link to bioconductor.
• set-enrichment methods - Query omic(s) dataset(s) against predefined omic(s) set, to see which omic(s) sets are enriched in the datasets.
  • "Classical" GSEA
  • GAGE

Extra Tools

These are just here for the record. These are more advanced, but aren't actually necessarily better for various reasons.

• Reactome - web app that does some exploratory visualization. Has a fairly major drawback that it only accepts, at most, a table of AUC values, without group information (i.e. no between-group comparisons are possible).
• Biocyc ptools - This has a free number of logins, after which a paid license is required. Unfortunately, not a free academic license for Melbourne Uni staff.

xcms Parameter Selection

• LCMS - from W4M notes.

Normalization

Long, introductory text to normalization here.