Improving sample classification by harnessing the potential of ¹H-NMR signal chemical shifts.

ABSTRACT: NMR spectroscopy is a technology that is widely used in metabolomic studies. The information that these studies most commonly use from NMR spectra is the metabolite concentration. However, as well as concentration, pH and ionic strength information are also made available by the chemical shift of metabolite signals. This information is typically not used even though it can enhance sample discrimination, since many conditions show pH or ionic imbalance. Here, we demonstrate how chemical shift information can be used to improve the quality of the discrimination between case and control samples in three public datasets of different human matrices. In two of these datasets, chemical shift information helped to provide an AUROC value higher than 0.9 during sample classification. In the other da

SUBMITTER: Canueto D 

PROVIDER: S-EPMC6082897 | biostudies-literature | 2018 Aug

REPOSITORIES: biostudies-literature