A Universal Framework for Blood Ionome Extraction and Intelligent Quality Control in ¹H NMR Metabolomics

Nuclear magnetic resonance (NMR) spectroscopy is a cornerstone of metabolomics and clinical bioanalysis; yet, its routine clinical adoption is limited by challenges in quality control (QC) and spectral interpretation. In blood-based NMR studies, ethylenediaminetetraacetic acid (EDTA) and its metal complexes are a major source of spectral interference, obscuring endogenous metabolites while increasing the analytical complexity. Here, we introduce a fully automated cheminformatics platform that enables EDTA-aware QC and the quantitative extraction of metal ion information while preserving metabolic integrity. Using large plasma and serum cohorts acquired at 500 and 600 MHz, we establish a multicenter-validated workflow for accurate quantification of calcium (Ca²⁺), magnesium (Mg²⁺), and zinc (Zn²⁺) directly from routine ¹H NMR spectra (R² > 0.9 versus clinical and spectrometric assays). The platform harmonizes data sets across instruments and cohorts through automated baseline reconstruction. Application to a large breast cancer cohort reveals a stage-dependent decline in circulating Zn²⁺, consistent with disrupted zinc homeostasis in invasive disease. This work bridges metabolomics and metallomics, advancing the translational utility of NMR in precision medicine.