Deep Learning Methodology for Obtaining Ultraclean Pure Shift Proton Nuclear Magnetic Resonance Spectra.

Nuclear magnetic resonance (NMR) is one of the most powerful analytical techniques. In order to obtain high-quality NMR spectra, a real-time Zangger-Sterk (ZS) pulse sequence is employed to collect low-quality pure shift NMR data with high efficiency. Then, a neural network named AC-ResNet and a loss function named SM-CDMANE are developed to train a network model. The model with excellent abilities of suppressing noise, reducing line widths, discerning peaks, and removing artifacts is utilized to process the acquired NMR data. The processed spectra with noise and artifact suppression and small line widths are ultraclean and high-resolution. Peaks overlapped heavily can be resolved. Weak peaks, even hidden in the noise, can be discerned from noise. Artifacts, even as high as spectral peaks, can be removed completely while not suppressing peaks. Eliminating perfectly noise and artifacts and smoothing baseline make spectra ultraclean. The proposed methodology would greatly promote various NMR applications.

Unambiguous and accurate measurement of scalar coupling constants through a selective refocusing NMR experiment.

Scalar coupling plays an important role in the analysis of molecular structure and dynamics. A great number of nuclear magnetic resonance (NMR) selective refocusing experiments, such as 2D G-SERF and PSYCHEDELIC, were developed to extract scalar coupling constants involving a selected proton from overlapped spectra. However, intense axial peaks occur in this type of experiments, leading to possible ambiguity in the assignment of spectral peaks and subsequent accurate measurement of 1H-1H scalar coupling constants. Here, a method based on selective coherence transfer and PSYCHEDELIC module is designed to acquire absorption-mode selective refocusing spectrum while suppressing intense axial peaks. Therefore, unambiguous and accurate measurement of scalar coupling constants involving the selectively excited proton can be achieved. The performances of the proposed method are demonstrated on several samples.

A simple data post-processing method for axial peaks free 2D PSYCHEDELIC NMR spectra.

Homonuclear scalar coupling plays an important role in the elucidation of molecular structure and dynamics. However, complex multiplets due to 1H-1H scalar coupling splittings complicate the assignment of peaks in overcrowded spectral regions. Although many methods focusing on disentangling couplings have been proposed in recent years, some defects like intense axial peaks and dispersive components still exist. Herein, a simple data post-processing method based on the interleaved acquisition mode PSYCHEDELIC (Pure Shift Yielded by CHirp Excitation to DELiver Individual Couplings) is designed to acquire absorption-mode 2D J spectrum while eradicating axial peaks. This approach provides a high resolution and pure absorptive spectrum, permitting unambiguous and accurate measurement of scalar coupling constants involving a given proton.

Improvement in Signal-to-Noise Ratio of Liquid-State NMR Spectroscopy via a Deep Neural Network DN-Unet.

Nuclear magnetic resonance (NMR) is one of the most powerful analytical tools and is extensively applied in many fields. However, compared to other spectroscopic techniques, NMR has lower sensitivity, impeding its wider applications. Using data postprocessing techniques to increase the NMR spectral signal-to-noise ratio (SNR) is a relatively simple and cost-effective method. In this work, a deep neural network, termed as DN-Unet, is devised to suppress noise in liquid-state NMR spectra to enhance SNR. It combines structures of encoder-decoder and convolutional neural network. Different from traditional deep learning training strategy, M-to-S strategy is developed to enhance DN-Unet capability that multiple noisy spectra (inputs) correspond to a same single noiseless spectrum (label) in the training stage. The trained 1D model can be used for denoising not only 1D but also high dimension spectra, further improving DN-Unet's performance. 1D, 2D, and 3D NMR spectra were utilized to evaluate DN-Unet performance. The results suggest that DN-Unet provides larger than 200-fold increase in SNR with weak peaks hidden in noise perfectly recovered and spurious peaks suppressed well. Since DN-Unet developed here to increase SNR is based on data postprocessing, it is universal for a variety of samples and NMR platforms. The great SNR enhancement and extreme excellence in differentiating signal and noise would greatly promote various liquid-state NMR applications.

Fully Exploiting the Power of 2D NMR J-Resolved Spectroscopy.

Nuclear magnetic resonance (NMR) spectroscopy is a powerful analytical tool that enables one to study molecular properties and interactions. Homonuclear couplings provide valuable structural information but are often difficult to disentangle in crowded 1H NMR spectra where complex multiplets and signal overlap commonly exist. Multidimensional NMR experiments push the power of NMR to a new level by providing better signal dispersion. Among them, 2D J-resolved spectroscopy is widely used for multiplet analysis and the measurement of scalar coupling constants. Here, we present a new 2D J-resolved method, CASCADE, through which easier multiplet analysis and unambiguous measurement of specific coupling constants can be achieved at the same time, fully exploiting the power of 2D J-resolved spectroscopy. It is expected that this method may replace a conventional 2D J experiment in many cases, facilitating structural and configurational studies as well as chemical and biological analyses.

PE-SERF: A sensitivity-improved experiment to measure JHH in crowded spectra.

Aiming at facilitating the analysis of molecular structure, the gradient-encoded selective refocusing methods (G-SERF) and a great number of its variants for measuring proton-proton coupling constants have been proposed. However, the sensitivity is an issue in the 2D gradient-encoded experiments, because the signal intensity is determined by the slice thickness of the sample that depends on encoding gradient and the bandwidth of selective pulses which is limited by the smallest chemical shift difference of any two coupled protons. Here, we present a method dubbed PE-SERF (perfect echo selective refocusing) which can determine all JHH values involving a selected proton with improved sensitivity compared to original G-SERF experiment. The modules of perfect echo involving selective pulses and gradient-encoded selective refocusing are combined in the method, so that the unwanted J couplings arising from coupled spin pairs in the same sample slice would be nullified. In this way, instead of single proton, a pair of coupled protons is allowed to share a sample slice, and thus the slice thickness can be increased and the spectral sensitivity can be improved. The performance of the method is demonstrated by experiments on quinine and strychnine.