Evaluation of analytical, experimental and biological variability in a UPLC-MS liver metabolic profiling study

摘要

Triplicate injections clustered in PCA scores plots, indicating a low analytical variability compared to biological variation. For aqueous extracts, experimental variability could not be distinguished from analytical variability and the 9 injections from each rat (3 injections of 3 single portions) clustered together. For organic extracts, experimental variability was higher than analytical variability, resulting in 3 groups ((velence)portions) of 3 points (replicated injections) for most rats. This may be explained by the non homogeneity of liver sections in terms of lipid content. Separation between rats was seen though, and treatment groups were clearly separated. This study allowed us to investigate the reliability of our extraction protocol and UPLCMS methods for untargeted metabolic profiling of liver samples. Following these results, the protocol assessed here has been applied to the complete set of rat liver samples from this toxicological study (36 rats), as well as to other metabolomic studies involving liver metabolic profiling by UPLC-MS.