Detection of Exchangeable Protons in NMR Metabolomic Analysis Using AI-Designed Water Irradiation Devoid Pulses

Abstract

¹H NMR spectroscopy has enabled the quantitative profiling of metabolites in various biofluids, emerging as a possible diagnostic tool for metabolic disorders and other diseases. To boost the signal-to-noise ratio and detect proton resonances near the water signal, current ¹H NMR experiments require solvent suppression schemes (e.g., presaturation, jump-and-return, WATERGATE, excitation sculpting, etc.). Unfortunately, these techniques affect the quantitative assessment of analytes containing exchangeable protons. To address this issue, we introduce two new one-dimensional (1D) ¹H NMR techniques that eliminate the water signal, preserving the intensities of exchangeable protons. Using GENETICS-AI, a software that combines an evolutionary algorithm and artificial intelligence, we tailored new water irradiation devoid (WADE) pulses and optimized the 1D ¹H NOESY sequence for metabolomic analysis. When applied to human urine samples, kidney tissue extract, and plasma, the WADE technique allowed for accurate measurement of typical metabolites and direct quantification of urea, which is usually challenging to measure using standard NMR experiments. We anticipate that these new NMR techniques will significantly improve the accuracy and reliability of metabolite quantitative assessment for a wide range of biological fluids.