NMR signal processing, prediction, and structure verification with machine learning techniques

Abstract

Machine learning (ML) methods have been present in the field of NMR since decades, but it has experienced a tremendous growth in the last few years, especially thanks to the emergence of deep learning (DL) techniques taking advantage of the increased amounts of data and available computer power. These algorithms are successfully employed for classification, regression, clustering, or dimensionality reduction tasks of large data sets and have been intensively applied in different areas of NMR including metabonomics, clinical diagnosis, or relaxometry. In this article, we concentrate on the various applications of ML/DL in the areas of NMR signal processing and analysis of small molecules, including automatic structure verification and prediction of NMR observables in solution. K E Y W O R D S

A data‐oriented approach to making new molecules as a student experiment: artificial intelligence‐enabling FAIR publication of NMR data for organic esters

“…[24] These are just some examples; a recent review is available. [25] It is clear that NMR data of all levels should be available in a systematic and predictable fashion for exploitation by AI techniques. In the present article, we describe an alternative approach to making data available by publication and which requires an infrastructure including a general metadata registration agency, [26] which is complementary to that for used journal publishing [27] and access to a data publication repository, which can capture a metadata record rich enough to carry at least information relevant to NMR data.…”

Perspective on integrating machine learning into computational chemistry and materials science

“…Another application where ML shows promise is the automated interpretation of nuclear magnetic resonance spectra with respect to atomic structure, which typically relies heavily on experience. 284 However, ML can also be used to leverage information contained in large collections of scientific data. The majority of chemical knowledge is collected in the form of publications.…”

Deep metabolome: Applications of deep learning in metabolomics

“…Although there are multiple steps of data processing and analysis, most applications of DL in metabolomics were for signal processing. This may be in part due to the large data requirements for DL for which simulation or synthetic data creation have been proposed [39] . Hansen (2019) proposed to use DNN to reconstruct non-uniformly sampling (NUS) NMR spectra.…”