Matrix effects in untargeted LC-MS metabolomics: From creation to compensation with post-column infusion of standards

• Post-column infusion of standards (PCIS) is used to address matrix effects • Strategy is presented to select an optimal PCIS using artificial matrix infusion • New scoring system balances relative and absolute matrix effect • 89% agreement in PCIS selection using artificial vs. biological matrix effect • The method expands PCIS use to untargeted LC-MS and enhances data accuracy Matrix effect is a well-known issue affecting accuracy and repeatability in metabolomics studies using liquid chromatography-electrospray ionization-mass spectrometry (LC-ESI-MS). Post-column infusion of standards (PCIS) is a promising strategy to monitor and correct matrix effect but has been rarely reported in untargeted metabolomics. The major challenges lie in selecting appropriate PCISs and identifying the most suitable PCIS to correct the matrix effect experienced by each feature. In this study, we aim to present a method for selecting suitable PCISs for matrix effect compensation based on the artificial matrix effect (ME art ) created by post-column infusion of compounds that disrupt the ESI process. Our hypothesis is that the suitable PCIS for a given analyte can be identified by comparing the PCISs’ ability in ME art compensation. We evaluated this approach using 19 stable-isotopically labeled (SIL) standards spiked in plasma, urine, and feces. PCISs selected based on ME art were compared to those selected by biological matrix effect (ME bio ), with 17 out of 19 SIL standards (89%) showing consistent PCIS selection, demonstrating the effectiveness of ME art in identifying suitable PCISs. Applying ME art -selected PCISs to correct for the ME bio resulted in improved ME bio for most of the SILs affected by matrix effect and maintained ME bio for those experiencing no matrix effect. We demonstrated the efficacy of ME art in selecting suitable PCISs for ME bio correction within an LC-PCIS-MS method. Importantly, since ME art can be assessed for any detected feature, its application holds great potential for identifying suitable PCISs for matrix effect correction in untargeted metabolomics.