Magnetstein: Web application for quantitative NMR mixture analysis

In this paper, we present a novel web application for the quantitative analysis of nuclear magnetic resonance (NMR) spectra of mixtures. At its core is the previously published Magnetstein algorithm, which solves the problem of quantification, i.e., determines the relative amounts of components in a mixture. Our innovative approach to this problem, based on the mathematical concept of the Wasserstein distance, makes the algorithm robust to common challenges such as distorted line shapes, spectral overlap, peak shifts, and low resolution. Thus far, Magnetstein has been available solely in an open-source Python package. Now, by creating the web service with a graphical user interface, we make our software available to the broader scientific community. The usage of our web application requires no knowledge of Python or other programming languages and is open and free of charge for academics. • We present a web application for quantitative analysis of NMR spectra of mixtures. • The processing is based on the mathematical concept of the Wasserstein distance. • Our method handles spectral overlap, shape distortions, peak shifts, low resolution.