Impact of the pre-examination phase on multicenter metabolomic studies
Veronica Ghini, Peter M. Abuja, Ozren Polašek, Łukasz Kozera, Päivi Laiho, Gabriele Anton, Marie Zins, Jānis Kloviņš, Andres Metspalu, H.‐Erich Wichmann, Christian Gieger, Claudio Luchinat, Kurt Zatloukal, Paola Turano
Abstract
The development of metabolomics in clinical applications has been limited by the lack of validation in large multicenter studies. Large population cohorts and their biobanks are a valuable resource for acquiring insights into molecular disease mechanisms. Nevertheless, most of their collections are not tailored for metabolomics and have been created without specific attention to the pre-analytical requirements for high-quality metabolome assessment. Thus, comparing samples obtained by different pre-analytical procedures remains a major challenge. Here, 1H NMR-based analyses are used to demonstrate how human serum and plasma samples collected with different operating procedures within several large European cohort studies from the Biobanking and Biomolecular Resources Infrastructure – Large Prospective Cohorts (BBMRI-LPC) consortium can be easily revealed by supervised multivariate statistical analyses at the initial stages of the process, to avoid biases in the downstream analysis. The inter-biobank differences are discussed in terms of deviations from the validated CEN/TS 16945:2016 / ISO 23118:2021 norms. It clearly emerges that biobanks must adhere to the evidence-based guidelines in order to support wider-scale application of metabolomics in biomedicine, and that NMR spectroscopy is informative in comparing the quality of different sample sources in multi cohort/center studies.