Improving data quality in liquid chromatography-mass spectrometry metabolomics of human urine

A B S T R A C T Signal variation is a common drawback in untargeted metabolomics using liquid chromatography-mass spectrometry (LC-MS), mainly due to the complexity of biological matrices and reduced sample prepara-tion, which results in the accumulation of sample components in the column and the ion source. Here we propose a simple, easy to implement approach to improve data quality in untargeted metabolomics by LC-MS. This approach involves the use of a divert valve to direct the column effluent to waste at the be-ginning of the chromatographic run and during column cleanup and equilibration, in combination with longer column cleanups in between injections. Our approach was tested using urine samples collected from patients after renal transplantation. Analytical responses were contrasted before and after introduc-ing these modifications by analyzing a batch of untargeted metabolomics data. A significant improvement in peak area repeatability was observed for the quality controls, with relative standard deviations (RSDs) for several metabolites decreasing from-60% to-10% when our approach was introduced. Similarly, RSDs of peak areas for internal standards improved from-40% to-10%. Furthermore, calibrant solutions were more consistent after introducing these modifications when comparing peak areas of solutions in-jected at the beginning and the end of each analytical sequence. Therefore, we recommend the use of a divert valve and extended column cleanup as a powerful strategy to improve data quality in untar-geted metabolomics, especially for very complex types of samples where minimum sample preparation is required, such as in this untargeted metabolomics study with urine from renal transplanted patients. (c) 2021 Elsevier B.V. All rights reserved. (AU)