Table 1.
Summary of measurement technologies and analysis techniques discussed in this review with selected example references
| Measurement technologies | Description |
|---|---|
| Mass spectrometry | Rapid detection of low-concentration metabolites103 |
| GC-MS | Separation of volatile metabolites104 |
| LC-MS | Separation of non-volatile metabolites, broad scope104 |
| Direct infusion | Fast broad coverage of metabolites105 |
| High-throughput NMR | Complementary measurement technology; precise concentration measurement106 |
| Metabolic flux imaging | In situ measurement of reaction rates in patients; non-invasive and non-destructive81 |
| Analysis techniques | Description |
| Metabolic association studies | Direct analogue of GWAS studies; testing of metabolites for association with phenotype24 |
| Gaussian graphical modelling | Inference and reconstruction of metabolic pathways where reactions are unknown27 |
| Pathway analysis | Test for enrichment of sets of functionally related entities associated with phenotype33 |
| Gene set enrichment analysis | Gene sets sourced from databases and ontologies (e.g. Gene Ontology)30 |
| Metabolic set enrichment analysis | Metabolite sets sourced from databases (e.g. KEGG database)29 |
| Metabolomic GWAS | Finding single nucleotide polymorphism s (SNPs) correlated with metabolic markers; GWAS with metabolite as trait42 |
| Classic mendelian randomization | Determination of causal relationships between an exposure and outcome of interest using SNP as instrument55 |
| Two-step MR | As for classic MR, but enables the testing of intermediate phenotypes that may confound the instrument107 |
| Metabolite association with co-expression networks | Association of metabolite measurements with systems of genes that have similar expression behaviour76 |
| Metabolite ratios | Association of ratios of metabolites, used as proxies for reaction rates, with a phenotype45 |
| Genome-scale model simulation | Simulation of known reactions incorporating genetic variation97 |