Untargeted metabolomics HRMS data processing using regions of interest and multivariate curve resolution approaches to unveil health-to-disease transition

• New workflow based on HRMS and ROIMCR for human plasma metabolomic profiling. • Positive and negative data processed simultaneously without requiring time alignment. • Chronic kidney disease selected as case study. • Multivariate analysis evidenced samples clustering according to health status. • Identification of health-to-disease transition biomarkers, including early markers. Early diagnosis has the potential to prevent or minimize the progression of diseases. The discovery of biomarkers that permit to pinpoint the health-to-disease transition is thus of crucial importance to improve diagnostic efficiency and treatment. Over the last years, untargeted metabolomic analysis of biomatrices has become a valuable tool to identify biomarkers of health status. However, it is still a recent research field with many challenges to address, namely the implementation of data treatment strategies that allow to deal efficiently with the highly complex data sets generated, without losing accuracy or relevant information. This work pursued the application of the chemometrics method Regions of Interest-Multivariate Curve Resolution (ROIMCR) to the MS-based metabolomic profiling of plasma samples aiming at the identification of chronic kidney disease (CKD) biomarkers. ROIMCR successfully resolved the plasma profiles of three groups of individuals: healthy controls, intermediate stage (pre-dialysis) and end-stage (in dialysis) CKD patients. Positive (MS1+) and negative (MS1−) data were processed simultaneously without requiring previous time alignment, resulting in time-effective analysis with increased metabolite coverage and identification. Multivariate analysis (PCA, PLS-DA, ASCA) revealed that samples were clustered according to health status and permitted to determine the most influential metabolites in differentiating groups. Metabolites identification evidenced recognized biomarkers of CKD, validating the proposed approach, and potential new indicators of disease onset and progression. A new analytical workflow involving instrumental analysis by UHPLC-HRMS and data treatment by chemometrics method ROIMCR, followed by multivariate analysis, is available for plasma metabolomic profiling. This platform can be easily adapted to other biomatrices and diseases, and applied to profile the same individual along time, fostering the development of personalized medicine.