Virtual Homonuclear Decoupling in Direct Detection Nuclear Magnetic Resonance Experiments Using Deep Neural Networks

WPR-Net: A Deep Learning Protocol for Highly Accelerated NMR Spectroscopy with Faithful Weak Peak Reconstruction

Multidimensional NMR spectroscopy contains a large amount of molecular-level species and structure information, which is of great significance in various disciplines; however, it is unfortunately limited by lengthy acquisition times. Undersampling signals accompanied by spectral reconstruction provide a powerful and efficient way to accelerate its implementation. However, the accurate reconstruction of weak peaks remains a crucial issue to compromise the reconstruction performance. In this work, we introduce a deep learning architecture for highly accelerated NMR spectroscopy along with the reliable reconstruction of weak peaks. This deep learning protocol allows one to eliminate undersampled artifacts and reconstruct high-quality multidimensional NMR spectroscopy signals, even under the conditions of highly sparse sampling density or in the presence of severe noise. Therefore, this study provides a powerful tool for fast multidimensional NMR spectroscopy and presents meaningful application prospects toward broader chemical and biological applications.

Deep learning and its applications in nuclear magnetic resonance spectroscopy

Nuclear Magnetic Resonance (NMR), as an advanced technology, has widespread applications in various fields like chemistry, biology, and medicine. However, issues such as long acquisition times for multidimensional spectra and low sensitivity limit the broader application of NMR. Traditional algorithms aim to address these issues but have limitations in speed and accuracy. Deep Learning (DL), a branch of Artificial Intelligence (AI) technology, has shown remarkable success in many fields including NMR. This paper presents an overview of the basics of DL and current applications of DL in NMR, highlights existing challenges, and suggests potential directions for improvement.

A combined NMR and deep neural network approach for enhancing the spectral resolution of aromatic side chains in proteins

Nuclear magnetic resonance (NMR) spectroscopy is an important technique for deriving the dynamics and interactions of macromolecules; however, characterizations of aromatic residues in proteins still pose a challenge. Here, we present a deep neural network (DNN), which transforms NMR spectra recorded on simple uniformly 13C-labeled samples to yield high-quality 1H-13C correlation maps of aromatic side chains. Key to the success of the DNN is the design of NMR experiments that produce data with unique features to aid the DNN produce high-resolution spectra. The methodology was validated experimentally on protein samples ranging from 7 to 40 kDa in size, where it accurately reconstructed multidimensional aromatic 1H-13C correlation maps, to facilitate 1H-13C chemical shift assignments and to quantify kinetics. More generally, we believe that the strategy of designing new NMR experiments in combination with customized DNNs represents a substantial advance that will have a major impact on the study of molecules using NMR in the years to come.

Neural net analysis of NMR spectra from strongly-coupled spin systems

Extracting parameters such as chemical shifts and coupling constants from proton NMR spectra is often a first step in using spectra for compound identification and structure determination. This can become challenging when scalar couplings between protons are comparable in size to chemical shift differences (strongly coupled), as is often the case with low-field (bench top) spectrometers. Here we explore the potential utility of AI methods, in particular neural networks, for extracting parameters from low-field spectra. Rather than seeking large experimental sets of spectra for training a network, we chose quantum mechanical simulation of sets, something that is possible with modern software packages and computer resources. We show that application of a network trained on 2-D J-resolved spectra and applied to a spectrum of iduronic acid, shows some promise, but also meets with some obstacles. We suggest that these may be overcome with improved pulse sequences and more extensive simulations.

Beyond traditional magnetic resonance processing with artificial intelligence

Smart signal processing approaches using Artificial Intelligence are gaining momentum in NMR applications. In this study, we demonstrate that AI offers new opportunities beyond tasks addressed by traditional techniques. We developed and trained artificial neural networks to solve three problems that until now were deemed "impossible": quadrature detection using only Echo (or Anti-Echo) modulation from the traditional Echo/Anti-Echo scheme; accessing uncertainty of signal intensity at each point in a spectrum processed by any given method; and defining a reference-free score for quantitative access of NMR spectrum quality. Our findings highlight the potential of AI techniques to revolutionize NMR processing and analysis.

Protein NMR assignment by isotope pattern recognition

The current standard method for amino acid signal identification in protein NMR spectra is sequential assignment using triple-resonance experiments. Good software and elaborate heuristics exist, but the process remains laboriously manual. Machine learning does help, but its training databases need millions of samples that cover all relevant physics and every kind of instrumental artifact. In this communication, we offer a solution to this problem. We propose polyadic decompositions to store millions of simulated three-dimensional NMR spectra, on-the-fly generation of artifacts during training, a probabilistic way to incorporate prior and posterior information, and integration with the industry standard CcpNmr software framework. The resulting neural nets take [1H,13C] slices of mixed pyruvate-labeled HNCA spectra (different CA signal shapes for different residue types) and return an amino acid probability table. In combination with primary sequence information, backbones of common proteins (GB1, MBP, and INMT) are rapidly assigned from just the HNCA spectrum.

Fast Pure Shift NMR Spectroscopy Using Attention-Assisted Deep Neural Network

Pure shift NMR spectroscopy enables the robust probing on molecular structure and dynamics, benefiting from great resolution enhancements. Despite extensive application landscapes in various branches of chemistry, the long experimental times induced by the additional time dimension generally hinder its further developments and practical deployments, especially for multi-dimensional pure shift NMR. Herein, this study proposes and implements the fast, reliable, and robust reconstruction for accelerated pure shift NMR spectroscopy with lightweight attention-assisted deep neural network. This deep learning protocol allows one to regain high-resolution signals and suppress undersampling artifacts, as well as furnish high-fidelity signal intensities along with the accelerated pure shift acquisition, benefitting from the introduction of the attention mechanism to highlight the spectral feature and information of interest. Extensive results of simulated and experimental NMR data demonstrate that this attention-assisted deep learning protocol enables the effective recovery of weak signals that are almost drown in the serious undersampling artifacts, and the distinction and recognition of close chemical shifts even though using merely 5.4% data, highlighting its huge potentials on fast pure shift NMR spectroscopy. As a result, this study affords a promising paradigm for the AI-assisted NMR protocols toward broader applications in chemistry, biology, materials, and life sciences, and among others.

Solution-state methyl NMR spectroscopy of large non-deuterated proteins enabled by deep neural networks

Methyl-TROSY nuclear magnetic resonance (NMR) spectroscopy is a powerful technique for characterising large biomolecules in solution. However, preparing samples for these experiments is demanding and entails deuteration, limiting its use. Here we demonstrate that NMR spectra recorded on protonated, uniformly 13C labelled samples can be processed using deep neural networks to yield spectra that are of similar quality to typical deuterated methyl-TROSY spectra, potentially providing information for proteins that cannot be produced in bacterial systems. We validate the methodology experimentally on three proteins with molecular weights in the range 42-360 kDa. We further demonstrate the applicability of our methodology to 3D NOESY spectra of Escherichia coli Malate Synthase G (81 kDa), where observed NOE cross-peaks are in good agreement with the available structure. The method represents an advance in the field of using deep learning to analyse complex magnetic resonance data and could have an impact on the study of large biomolecules in years to come.

Optimal 13C NMR investigation of intrinsically disordered proteins at 1.2 GHz

Nuclear magnetic resonance (NMR) spectroscopy is a powerful technique for characterizing biomolecules such as proteins and nucleic acids at atomic resolution. Increased magnetic field strengths drive progress in biomolecular NMR applications, leading to improved performance, e.g., higher resolution. A new class of NMR spectrometers with a 28.2 T magnetic field (1.2 GHz 1H frequency) has been commercially available since the end of 2019. The availability of ultra-high-field NMR instrumentation makes it possible to investigate more complex systems using NMR. This is especially true for highly flexible intrinsically disordered proteins (IDPs) and highly flexible regions (IDRs) of complex multidomain proteins. Indeed, the investigation of these proteins is frequently hampered by the crowding of NMR spectra. The advantages, however, are accompanied by challenges that the user must overcome when conducting experiments at such a high field (e.g., large spectral widths, radio frequency bandwidth, performance of decoupling schemes). This protocol presents strategies and tricks for optimising high-field NMR experiments for IDPs/IDRs based on the analysis of the relaxation properties of the investigated protein. The protocol, tested on three IDPs of different molecular weight and structural complexity, focuses on 13C-detected NMR at 1.2 GHz. A set of experiments, including some multiple receiver experiments, and tips to implement versions tailored for IDPs/IDRs are described. However, the general approach and most considerations can also be applied to experiments that acquire 1H or 15N nuclei and to experiments performed at lower field strengths.

Time-optimized protein NMR assignment with an integrative deep learning approach using AlphaFold and chemical shift prediction

Chemical shift assignment is vital for nuclear magnetic resonance (NMR)-based studies of protein structures, dynamics, and interactions, providing crucial atomic-level insight. However, obtaining chemical shift assignments is labor intensive and requires extensive measurement time. To address this limitation, we previously proposed ARTINA, a deep learning method for automatic assignment of two-dimensional (2D)-4D NMR spectra. Here, we present an integrative approach that combines ARTINA with AlphaFold and UCBShift, enabling chemical shift assignment with reduced experimental data, increased accuracy, and enhanced robustness for larger systems, as presented in a comprehensive study with more than 5000 automated assignment calculations on 89 proteins. We demonstrate that five 3D spectra yield more accurate assignments (92.59%) than pure ARTINA runs using all experimentally available NMR data (on average 10 3D spectra per protein, 91.37%), considerably reducing the required measurement time. We also showcase automated assignments of only 15N-labeled samples, and report improved assignment accuracy in larger synthetic systems of up to 500 residues.

Biomolecular NMR spectroscopy in the era of artificial intelligence

Biomolecular nuclear magnetic resonance (NMR) spectroscopy and artificial intelligence (AI) have a burgeoning synergy. Deep learning-based structural predictors have forever changed structural biology, yet these tools currently face limitations in accurately characterizing protein dynamics, allostery, and conformational heterogeneity. We begin by highlighting the unique abilities of biomolecular NMR spectroscopy to complement AI-based structural predictions toward addressing these knowledge gaps. We then highlight the direct integration of deep learning approaches into biomolecular NMR methods. AI-based tools can dramatically improve the acquisition and analysis of NMR spectra, enhancing the accuracy and reliability of NMR measurements, thus streamlining experimental processes. Additionally, deep learning enables the development of novel types of NMR experiments that were previously unattainable, expanding the scope and potential of biomolecular NMR spectroscopy. Ultimately, a combination of AI and NMR promises to further revolutionize structural biology on several levels, advance our understanding of complex biomolecular systems, and accelerate drug discovery efforts.

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.

Pure Isotropic Proton NMR Spectra in Solids using Deep Learning

The resolution of proton solid-state NMR spectra is usually limited by broadening arising from dipolar interactions between spins. Magic-angle spinning alleviates this broadening by inducing coherent averaging. However, even the highest spinning rates experimentally accessible today are not able to completely remove dipolar interactions. Here, we introduce a deep learning approach to determine pure isotropic proton spectra from a two-dimensional set of magic-angle spinning spectra acquired at different spinning rates. Applying the model to 8 organic solids yields high-resolution 1 H solid-state NMR spectra with isotropic linewidths in the 50-400 Hz range.

NMR spectrum reconstruction as a pattern recognition problem

A new deep neural network based on the WaveNet architecture (WNN) is presented, which is designed to grasp specific patterns in the NMR spectra. When trained at a fixed non-uniform sampling (NUS) schedule, the WNN benefits from pattern recognition of the corresponding point spread function (PSF) pattern produced by each spectral peak resulting in the highest quality and robust reconstruction of the NUS spectra as demonstrated in simulations and exemplified in this work on 2D ¹H-¹⁵N correlation spectra of three representative globular proteins with different sizes: Ubiquitin (8.6 kDa), Azurin (14 kDa), and Malt1 (44 kDa). The pattern recognition by WNN is also demonstrated for successful virtual homo-decoupling in a 2D methyl ¹H-¹³C - HMQC spectrum of MALT1. We demonstrate using WNN that prior knowledge about the NUS schedule, which so far was not been fully exploited, can be used for designing new powerful NMR processing techniques that surpass the existing algorithmic methods.

Less is more: A simple methyl-TROSY based pulse scheme offers improved sensitivity in applications to high molecular weight complexes

The HMQC pulse sequence and variants thereof have been exploited in studies of high molecular weight protein complexes, taking advantage of the fact that fast and slow relaxing magnetization components are sequestered along two distinct magnetization transfer pathways. Despite the simplicity of the HMQC scheme an even shorter version can be designed, based on elimination of the terminal refocusing period, as a further means of increasing signal. Here we present such an experiment, and show that significant sensitivity gains, in some cases by factors of two or more, are realized in studies of proteins varying in molecular masses from 100 kDa to 1 MDa.

Towards autonomous analysis of chemical exchange saturation transfer experiments using deep neural networks

Macromolecules often exchange between functional states on timescales that can be accessed with NMR spectroscopy and many NMR tools have been developed to characterise the kinetics and thermodynamics of the exchange processes, as well as the structure of the conformers that are involved. However, analysis of the NMR data that report on exchanging macromolecules often hinges on complex least-squares fitting procedures as well as human experience and intuition, which, in some cases, limits the widespread use of the methods. The applications of deep neural networks (DNNs) and artificial intelligence have increased significantly in the sciences, and recently, specifically, within the field of biomolecular NMR, where DNNs are now available for tasks such as the reconstruction of sparsely sampled spectra, peak picking, and virtual decoupling. Here we present a DNN for the analysis of chemical exchange saturation transfer (CEST) data reporting on two- or three-site chemical exchange involving sparse state lifetimes of between approximately 3-60 ms, the range most frequently observed via experiment. The work presented here focuses on the ¹H CEST class of methods that are further complicated, in relation to applications to other nuclei, by anti-phase features. The developed DNNs accurately predict the chemical shifts of nuclei in the exchanging species directly from anti-phase ¹H^N CEST profiles, along with an uncertainty associated with the predictions. The performance of the DNN was quantitatively assessed using both synthetic and experimental anti-phase CEST profiles. The assessments show that the DNN accurately determines chemical shifts and their associated uncertainties. The DNNs developed here do not contain any parameters for the end-user to adjust and the method therefore allows for autonomous analysis of complex NMR data that report on conformational exchange.

In-Cell Structural Biology by NMR: The Benefits of the Atomic Scale

In-cell structural biology aims at extracting structural information about proteins or nucleic acids in their native, cellular environment. This emerging field holds great promise and is already providing new facts and outlooks of interest at both fundamental and applied levels. NMR spectroscopy has important contributions on this stage: It brings information on a broad variety of nuclei at the atomic scale, which ensures its great versatility and uniqueness. Here, we detail the methods, the fundamental knowledge, and the applications in biomedical engineering related to in-cell structural biology by NMR. We finally propose a brief overview of the main other techniques in the field (EPR, smFRET, cryo-ET, etc.) to draw some advisable developments for in-cell NMR. In the era of large-scale screenings and deep learning, both accurate and qualitative experimental evidence are as essential as ever to understand the interior life of cells. In-cell structural biology by NMR spectroscopy can generate such a knowledge, and it does so at the atomic scale. This review is meant to deliver comprehensive but accessible information, with advanced technical details and reflections on the methods, the nature of the results, and the future of the field.