A New N-Channel Maximum Entropy Method in NMR for Automatic Reconstruction of “Decoupled Spectra” and J-Coupling Determination

A deep learning framework for multiplet splitting classification in 1H NMR

One-dimensional 1H Nuclear Magnetic Resonance (NMR) stands out as the quickest and simplest among various NMR experimental setups. Unfortunately, it suffers from lengthy annotation times and does not always have a clear and unique interpretation. From NMR discovery, efforts have been dedicated to introducing an automated approach to streamline the characterization of chemical compounds while ensuring consistency of the results across the scientific community. Nonetheless, this remains an ongoing challenge that has garnered renewed interest with the emergence of deep learning techniques. Here, we present MuSe Net, a novel supervised probabilistic deep learning framework that can emulate the tasks performed by an expert spectroscopist in annotating one-dimensional NMR spectra generated by small molecules. Considering only the spectrum, MuSe Net detects and classifies multiplets with up to four coupling constants for their splitting phenotype, providing a segmentation of the spectral range. We exploit uncertainty quantification to produce a confidence score to both assess classification reliability and to detect signals that do not fit into any other phenotype class. The results of the evaluation against 48 experimental 1H NMR spectra of small molecules annotated by experts demonstrate that MuSe Net can deal with anomalies and unclear signals while correctly classifying multiplets and detecting overlapping peaks.

Automatic classification of signal regions in 1H Nuclear Magnetic Resonance spectra

The identification and characterization of signal regions in Nuclear Magnetic Resonance (NMR) spectra is a challenging but crucial phase in the analysis and determination of complex chemical compounds. Here, we present a novel supervised deep learning approach to perform automatic detection and classification of multiplets in 1H NMR spectra. Our deep neural network was trained on a large number of synthetic spectra, with complete control over the features represented in the samples. We show that our model can detect signal regions effectively and minimize classification errors between different types of resonance patterns. We demonstrate that the network generalizes remarkably well on real experimental 1H NMR spectra.

Selective Amino Acid-Only in Vivo NMR: A Powerful Tool To Follow Stress Processes

In vivo NMR of small 13C-enriched aquatic organisms is developing as a powerful tool to detect and explain toxic stress at the biochemical level. Amino acids are a very important category of metabolites for stress detection as they are involved in the vast majority of stress response pathways. As such, they are a useful proxy for stress detection in general, which could then be a trigger for more in-depth analysis of the metabolome. 1H-13C heteronuclear single quantum coherence (HSQC) is commonly used to provide additional spectral dispersion in vivo and permit metabolite assignment. While some amino acids can be assigned from HSQC, spectral overlap makes monitoring them in vivo challenging. Here, an experiment typically used to study protein structures is adapted for the selective detection of amino acids inside living Daphnia magna (water fleas). All 20 common amino acids can be selectively detected in both extracts and in vivo. By monitoring bisphenol-A exposure, the in vivo amino acid-only approach identified larger fluxes in a greater number of amino acids when compared to published works using extracts from whole organism homogenates. This suggests that amino acid-only NMR of living organisms may be a very sensitive tool in the detection of stress in vivo and is highly complementary to more traditional metabolomics-based methods. The ability of selective NMR experiments to help researchers to "look inside" living organisms and only detect specific molecules of interest is quite profound and paves the way for the future development of additional targeted experiments for in vivo research and monitoring.

CONNJUR Workflow Builder: a software integration environment for spectral reconstruction

CONNJUR Workflow Builder (WB) is an open-source software integration environment that leverages existing spectral reconstruction tools to create a synergistic, coherent platform for converting biomolecular NMR data from the time domain to the frequency domain. WB provides data integration of primary data and metadata using a relational database, and includes a library of pre-built workflows for processing time domain data. WB simplifies maximum entropy reconstruction, facilitating the processing of non-uniformly sampled time domain data. As will be shown in the paper, the unique features of WB provide it with novel abilities to enhance the quality, accuracy, and fidelity of the spectral reconstruction process. WB also provides features which promote collaboration, education, parameterization, and non-uniform data sets along with processing integrated with the Rowland NMR Toolkit (RNMRTK) and NMRPipe software packages. WB is available free of charge in perpetuity, dual-licensed under the MIT and GPL open source licenses.

Modeling of NMR processing, toward efficient unattended processing of NMR experiments

Many alternative processing techniques have recently been proposed in the literature. Most of these techniques rely on specific acquisition protocols as well as on specific data processing techniques, the need for an efficient versatile and expandable NMR processing tool would be a particularly timely addition to the modern NMR spectroscopy laboratory. The work presented here consists in a modeling of the various possible NMR data processing approaches. This modeling presents a common working frame for most of the modern acquisition/processing protocols. Two different data modeling approaches are presented, strong modeling and weak modeling, depending whether the system under study or the measurement is modeled. The emphasis is placed on the weak modeling approach. This modeling is implemented in a computer program developed in python and called NPK standing (standing for NMR Processing Kernel), organized in four logical layers (i) mathematical kernel; (ii) elementary actions; (iii) processing phases; (iv) processing strategies. This organisation, along with default values for most processing parameters allows the use of the program in an unattended manner, producing close to optimal spectra. Examples are shown for 1D and 2D processing, and liquid and solid NMR spectroscopy. NPK is available from the site: http://abcis.cbs.cnrs.fr/NPK.

Elimination of 13Calpha splitting in protein NMR spectra by deconvolution with maximum entropy reconstruction

Homonuclear 13C-13C couplings can significantly reduce the sensitivity and resolution of multidimensional NMR experiments. The most important of these couplings is the 13Calpha-13Cbeta coupling, and several different methods have been developed to eliminate its effect from spectra used for backbone assignment, including short or constant-time evolution periods, selectively labeled amino acids, and multiple-band decoupling sequences. In this communication we show that postacquisition deconvolution of the spectra with a maximum entropy algorithm can be superior to experimental decoupling. The method is very robust, does not introduce shifts of the resonance positions, and simplifies the measurement of the most important NMR experiments for protein backbone assignment.