NMR is one of the best analytical tools for the quantification of samples and identification of unknown molecules. It is also a near-universal detector if the concentration of the analyte is sufficient. Given these great strengths, why is NMR metabolomics lagging LC-MS metabolomics?
My answer is that both quantification and unknown identification are still extremely difficult in NMR metabolomics.
In this presentation I describe a new approach that our lab is taking to fully quantify 1D 1H NMR metabolomics data. We start by creating a metabolite fraction library and then perform time-domain spectral analysis using SAND (Wu et al., 2024). Individual Lorentzian peaks are then correlated across fractions into a metabolite basis set that contains all the known and unknown metabolites in a particular sample that can be detected by NMR. Next, we use BATMAN (Hao et al., 2014) with the metabolite basis set as prior knowledge to fit metabolomics profiling data to get accurate quantification of all detectable metabolites in the NMR spectrum.
Presented By: Art Edison, University of Georgia