Abstract
The aim of this review is to show the risks of artifact formation in metabolomics analyses. Metabolomics has developed in a major tool in system biology approaches to unravel the metabolic networks that are the basis of life. Presently TLC, LC-MS, GC-MS, MS-MS and nuclear magnetic resonance are applied to analyze the metabolome of all kind of biomaterials. These analytical methods require robust preanalytical protocols to extract the small molecules from the biomatrix. The quality of the metabolomics analyses depends on protocols for collecting and processing of the biomaterial, including the methods for drying, grinding and extraction. Also the final preparation of the samples for instrumental analysis is crucial for highly reproducible analyses. The risks of artifact formation in these steps are reviewed from the point of view of the commonly used solvents. Examples of various artifacts formed through chemical reactions between solvents or contaminations with functional groups in the analytes are discussed. These reactions involve, for example, the formation of esters, trans-esterifications, hemiacetal and acetal formation, N-oxidations, and the formation of carbinolamines. It concerns chemical reactions with hydroxyl-, aldehyde-, keto-, carboxyl-, ester-, and amine functional groups. In the analytical steps, artifacts in LC may come from the stationary phase or reactions of the eluent with analytes. Differences between the solvent of the injected sample and the LC-mobile phase may cause distortions of the retention of analytes. In all analytical methods, poorly soluble compounds will be in all samples at saturation level, thus hiding a potential marker function. Finally a full identification of compounds remains a major hurdle in metabolomics, it requires a full set of spectral data, including methods for confirming the absolute stereochemistry. The putative identifications found in supplemental data of many studies, unfortunately, often become "truly" identified compounds in papers citing these results. Proper validation of the protocols for preanalytical and analytical procedures is essential for reproducible analyses in metabolomics.