Prediction Of Chemical Shifts Research Articles

Accurate Prediction of 1H NMR Chemical Shifts of Small Molecules Using Machine Learning

NMR is widely considered the gold standard for organic compound structure determination. As such, NMR is routinely used in organic compound identification, drug metabolite characterization, natural product discovery, and the deconvolution of metabolite mixtures in biofluids (metabolomics and exposomics). In many cases, compound identification by NMR is achieved by matching measured NMR spectra to experimentally collected NMR spectral reference libraries. Unfortunately, the number of available experimental NMR reference spectra, especially for metabolomics, medical diagnostics, or drug-related studies, is quite small. This experimental gap could be filled by predicting NMR chemical shifts for known compounds using computational methods such as machine learning (ML). Here, we describe how a deep learning algorithm that is trained on a high-quality, “solvent-aware” experimental dataset can be used to predict 1H chemical shifts more accurately than any other known method. The new program, called PROSPRE (PROton Shift PREdictor) can accurately (mean absolute error of <0.10 ppm) predict 1H chemical shifts in water (at neutral pH), chloroform, dimethyl sulfoxide, and methanol from a user-submitted chemical structure. PROSPRE (pronounced “prosper”) has also been used to predict 1H chemical shifts for >600,000 molecules in many popular metabolomic, drug, and natural product databases.

Highly Accurate Prediction of NMR Chemical Shifts from Low-Level Quantum Mechanics Calculations Using Machine Learning.

Theoretical predictions of NMR chemical shifts from first-principles can greatly facilitate experimental interpretation and structure identification of molecules in gas, solution, and solid-state phases. However, accurate prediction of chemical shifts using the gold-standard coupled cluster with singles, doubles, and perturbative triple excitations [CCSD(T)] method with a complete basis set (CBS) can be prohibitively expensive. By contrast, machine learning (ML) methods offer inexpensive alternatives for chemical shift predictions but are hampered by generalization to molecules outside the original training set. Here, we propose several new ideas in ML of the chemical shift prediction for H, C, N, and O that first introduce a novel feature representation, based on the atomic chemical shielding tensors within a molecular environment using an inexpensive quantum mechanics (QM) method, and train it to predict NMR chemical shieldings of a high-level composite theory that approaches the accuracy of CCSD(T)/CBS. In addition, we train the ML model through a new progressive active learning workflow that reduces the total number of expensive high-level composite calculations required while allowing the model to continuously improve on unseen data. Furthermore, the algorithm provides an error estimation, signaling potential unreliability in predictions if the error is large. Finally, we introduce a novel approach to keep the rotational invariance of the features using tensor environment vectors (TEVs) that yields a ML model with the highest accuracy compared to a similar model using data augmentation. We illustrate the predictive capacity of the resulting inexpensive shift machine learning (iShiftML) models across several benchmarks, including unseen molecules in the NS372 data set, gas-phase experimental chemical shifts for small organic molecules, and much larger and more complex natural products in which we can accurately differentiate between subtle diastereomers based on chemical shift assignments.

Prediction of 19F NMR chemical shift by machine learning

Fluorine-19 (19F) is a nucleus of great importance in the field of Nuclear Magnetic Resonance (NMR) spectroscopy due to its high receptivity and wide chemical shift dispersion. 19F NMR plays crucial roles in both organic synthesis and biomedicine. Herein, a machine learning-based comprehensive 19F NMR chemical shift prediction model was established based on the experimental 19F NMR dataset from the book by Dolbier and the open NMR database nmrshiftdb2. Fluorine radical SMILES (Fr-SMILES) that reflected the fluorine chemical equivalence, was designed as the representation of fluorine in the molecule. Model trained with the graph convolution network (GCN) algorithm gave a low mean absolute error (MAE) of 3.636 ppm on the testing set. This model exhibits broad applicability and can effectively predict 19F NMR shifts for a wide range of organic fluorine molecules. We believe that the current work will provide a powerful tool for not only predicting 19F NMR shifts but also aiding in the analysis and identification of these shifts in diverse organic fluorine compounds. An online prediction platform was constructed based on the current model, which can be found at https://fluobase.cstspace.cn/fnmr.

The interplay of density functional selection and crystal structure for accurate NMR chemical shift predictions

Ab initio chemical shift prediction plays a central role in nuclear magnetic resonance (NMR) crystallography, and the accuracy with which chemical shifts can be predicted relative to experiment impacts the...

Impacts of basis sets, solvent models, and NMR methods on the accuracy of 1H and 13C chemical shift calculations for biaryls: a DFT study

Introduction: Biaryls are core structures composed of chiral ligands, organocatalysts, biologically active natural products and biopolymer lignins. In this study, the effects of basis sets, solvent models, and NMR methods on the accuracy of 1H/13C NMR chemical shift calculations for biaryl structures were evaluated. Methods: All calculations were performed using Gaussian09. The GIAO NMR results were observed and extracted using GaussView05. To reduce the systematic error of the calculations, linear regression analysis of the calculated chemical shifts versus the experimental shifts was performed. Results: The tested basis sets showed good 1H/13C results, with CMAE values as low as 0.0425 ppm and 1.09 ppm for 1H and 13C, respectively. The use of solvent models significantly increased the accuracy of the 1H chemical shift calculations. The GIAO method produced more accurate results than did the IGAIM and CSGT methods. Conclusion: This study recommends 6-31G(d,p) and DGDZVP basis sets, IEMPCM and CPCM solvent models, and GIAO NMR methods for the accurate prediction of 1H and 13C chemical shifts for biaryls, assisting in their full structural assignments.

The accuracy limit of chemical shift predictions for species in aqueous solution.

Interpreting NMR experiments benefits from first-principles predictions of chemical shifts. Reaching the accuracy limit of theory is relevant for unambiguous structural analysis and dissecting theoretical approximations. Since accurate chemical shift measurements are based on using internal reference compounds such as trimethylsilylpropanesulfonate (DSS), a detailed comparison of experimental with theoretical data requires simultaneous consideration of both target and reference species ensembles in the same solvent environment. Here we show that ab initio molecular dynamics simulations to generate liquid-state ensembles of target and reference compounds, including explicitly their short-range solvation environments and combined with quantum-mechanical solvation models, allows for predicting highly accurate 1H (∼0.1-0.5 ppm) and aliphatic 13C (∼1.5 ppm) chemical shifts for aqueous solutions of the model compounds trimethylamine N-oxide (TMAO) and N-methylacetamide (NMA), referenced to DSS without any system-specific adjustments. This encompasses the two peptide bond conformations of NMA identified by NMR. The results are used to derive a general-purpose guideline set for predictive NMR chemical shift calculations of NMA in the liquid state and to identify artifacts of force field models. Accurate predictions are only obtained if a sufficient number of explicit water molecules is included in the quantum-mechanical calculations, disproving a purely electrostatic model of the solvent effect on chemical shifts.

COLMARppm: A Web Server Tool for the Accurate and Rapid Prediction of 1H and 13C NMR Chemical Shifts of Organic Molecules and Metabolites.

Despite rapid progress in metabolomics research, a major bottleneck is the large number of metabolites whose chemical structures are unknown or whose spectra have not been deposited in metabolomics databases. Nuclear magnetic resonance (NMR) spectroscopy has a long history of elucidating chemical structures from experimentally measured 1H and 13C chemical shifts. One approach to characterizing the chemical structures of an unknown metabolite is to predict the 1H and 13C chemical shifts of candidate compounds (e.g., metabolites from the Human Metabolome Database (HMDB)) and compare them with chemical shifts of the unknown. However, accurate prediction of NMR chemical shifts in aqueous solution is challenging due to limitations of experimental chemical shift libraries and the high computational cost of quantum chemical methods. To improve NMR prediction accuracy and applicability, an empirical prediction strategy is introduced here to provide an accurately predicted chemical shift for organic molecules and metabolites within seconds. Unique features of COLMARppm include (i) the training library exclusively consisting of high quality NMR spectra measured under standard conditions in aqueous solution, (ii) utilization of NMR motif information, and (iii) leveraging of the improved prediction accuracy for the automated assignment of experimental chemical shifts for candidate structures. COLMARppm is demonstrated in terms of accuracy and speed for a set of 20 compounds taken from the HMDB for chemical shift prediction and resonance assignment. COLMARppm is applicable to a wide range of small molecules and can be directly incorporated into metabolomics workflows.

Prediction of 19F NMR chemical shifts for organic compounds with ORCA

Accurate assignment of 19F NMR has long been a challenge, and quantum chemical methods are possible solutions. Herein we reported a scaling method for the prediction of 19F NMR chemical shift with freely available ORCA program package. Performance of 31 DFT functionals coupled with 11 basis sets were evaluated and influence of geometry optimization was also studied with five functionals coupled with three basis sets. The significance of geometry was further examined through the execution of relaxed surface scans of seven flexible compounds, and averaged shieldings of obtained conformers yielded notable improvement of the correlation between calculated isotropic shielidings and experimental chemical shifts. Utilization of the best scaling factor obtained successfully assigned of fluorine atoms in multifluorinated molecules with different conformations. The method reported here was computationally inexpensive, easily available with acceptable accuracy.

NMR shift prediction from small data quantities

Prediction of chemical shift in NMR using machine learning methods is typically done with the maximum amount of data available to achieve the best results. In some cases, such large amounts of data are not available, e.g. for heteronuclei. We demonstrate a novel machine learning model that is able to achieve better results than other models for relevant datasets with comparatively low amounts of data. We show this by predicting ^{19}F and ^{13}C NMR chemical shifts of small molecules in specific solvents.Graphical

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.

Automated Fragmentation Quantum Mechanical Calculation of 15N and 13C Chemical Shifts in a Membrane Protein.

In this work, we developed an accurate and cost-effective automated fragmentation quantum mechanics/molecular mechanics (AF-QM/MM) method to calculate the chemical shifts of 15N and 13C of membrane proteins. The convergence of the AF-QM/MM method was tested using Krokinobacter eikastus rhodopsin 2 as a test case. When the distance threshold of the QM region is equal to or larger than 4.0 Å, the results of the AF-QM/MM calculations are close to convergence. In addition, the effects of selected density functionals, basis sets, and local chemical environment of target atoms on the chemical shift calculations were systematically investigated. Our results demonstrate that the predicted chemical shifts are more accurate when important environmental factors including cross-protomer interactions, lipid molecules, and solvent molecules are taken into consideration, especially for the 15N chemical shift prediction. Furthermore, with the presence of sodium ions in the environment, the chemical shift of residues, retinal, and retinal Schiff base are affected, which is consistent with the results of the solid-state nuclear magnetic resonance (NMR) experiment. Upon comparing the performance of various density functionals (namely, B3LYP, B3PW91, M06-2X, M06-L, mPW1PW91, OB95, and OPBE), the results show that mPW1PW91 is a suitable functional for the 15N and 13C chemical shift prediction of the membrane proteins. Meanwhile, we find that the improved accuracy of the 13Cβ chemical shift calculations can be achieved by the employment of the triple-ζ basis set. However, the employment of the triple-ζ basis set does not improve the accuracy of the 15N and 13Cα chemical shift calculations nor does the addition of a diffuse function improve the overall prediction accuracy of the chemical shifts. Our study also underscores that the AF-QM/MM method has significant advantages in predicting the chemical shifts of key ligands and nonstandard residues in membrane proteins than most widely used empirical models; therefore, it could be an accurate computational tool for chemical shift calculations on various types of biological systems.

MIM-ML: A Novel Quantum Chemical Fragment-Based Random Forest Model for Accurate Prediction of NMR Chemical Shifts of Nucleic Acids.

We developed a random forest machine learning (ML) model for the prediction of 1H and 13C NMR chemical shifts of nucleic acids. Our ML model is trained entirely on reproducing computed chemical shifts obtained previously on 10 nucleic acids using a Molecules-in-Molecules (MIM) fragment-based density functional theory (DFT) protocol including microsolvation effects. Our ML model includes structural descriptors as well as electronic descriptors from an inexpensive low-level semiempirical calculation (GFN2-xTB) and trained on a relatively small number of DFT chemical shifts (2080 1H chemical shifts and 1780 13C chemical shifts on the 10 nucleic acids). The ML model is then used to make chemical shift predictions on 8 new nucleic acids ranging in size from 600 to 900 atoms and compared directly to experimental data. Though no experimental data was used in the training, the performance of our model is excellent (mean absolute deviation of 0.34 ppm for 1H chemical shifts and 2.52 ppm for 13C chemical shifts for the test set), despite having some nonstandard structures. A simple analysis suggests that both structural and electronic descriptors are critical for achieving reliable predictions. This is the first attempt to combine ML from fragment-based DFT calculations to predict experimental chemical shifts accurately, making the MIM-ML model a valuable tool for NMR predictions of nucleic acids.

NMR chemical shift prediction and structural elucidation of linker-containing oligo- and polysaccharides using the computer program CASPER

Carbohydrate structures containing alkyl groups as aglycones are useful for investigating enzyme activity and glycan-protein interactions. Moreover, linker-containing oligosaccharides with a spacer group are commonly used to print glycan microarrays or to prepare protein-conjugates as vaccine candidates. The structural accuracy of these synthesized glycans are essential for interpretation of results from biological experiments in which the compounds have been used and NMR spectroscopy can unravel and confirm their structures. An approach for efficient 1H and 13C NMR chemical shift assignments employed a parallel NOAH-10 measurement followed by NMR spin-simulation to refine the 1H NMR chemical shifts, as exemplified for a disaccharide with an azidoethyl group as an aglycone, the NMR chemical shifts of which have been used to enhance the quality of CASPER (http://www.casper.organ.su.se/casper/). The CASPER program has been further developed to aid characterization of linker-containing oligo- and polysaccharides, either by chemical shift prediction for comparison to experimental NMR data or as structural investigation of synthesized glycans based on acquired unassigned NMR data. The ability of CASPER to elucidate structures of linker-containing oligosaccharides is demonstrated and comparisons to assigned or unassigned NMR data show the utility of CASPER in supporting a proposed oligosaccharide structure. Prediction of NMR chemical shifts of an oligosaccharide, corresponding to the repeating unit of an O-antigen polysaccharide, having a linker as an aglycone and a non-natural substituent derivative thereof are presented to exemplify the diversity of structures handled. Furthermore, NMR chemical shift predictions of synthesized polysaccharides, corresponding to bacterial polysaccharides, containing a linker are described showing that in addition to oligosaccharide structures also polysaccharide structures having an aglycone spacer group can be analyzed by CASPER.

Rapid and Automated Ab Initio Metabolite Collisional Cross Section Prediction from SMILES Input.

We implemented an ab initio CCS prediction workflow which incrementally refines generated structures using molecular mechanics, a deep learning potential, conformational clustering, and quantum mechanics (QM). Automating intermediate steps for a high performance computing (HPC) environment allows users to input the SMILES structure of small organic molecules and obtain a Boltzmann averaged collisional cross section (CCS) value as output. The CCS of a molecular species is a metric measured by ion mobility spectrometry (IMS) which can improve annotation of untargeted metabolomics experiments. We report only a minor drop in accuracy when we expedite the CCS calculation by replacing the QM geometry refinement step with a single-point energy calculation. Even though the workflow involves stochastic steps (i.e., conformation generation and clustering), the final CCS value was highly reproducible for multiple iterations on L-carnosine. Finally, we illustrate that the gas phase ensembles modeled for the workflow are intermediate files which can be used for the prediction of other properties such as aqueous phase nuclear magnetic resonance chemical shift prediction. The software is available at the following link: https://github.com/DasSusanta/snakemake_ccs.

Versa DB: Assisting 13C NMR and MS/MS Joint Data Annotation Through On‐Demand Databases

AbstractCompound identification in complex mixtures by NMR and MS is best achieved through experimental databases (DB) mining. Experimental DB frequently show limitations regarding their completeness, availability or data quality, thus making predicted database of increasing common use. Querying large databases may lead to select unlikely structure candidates. Two approaches to dereplication are thus possible: filtering of a large DB before search or scoring of the results after a large scale search. The present work relies on the former approach. As far as we know, nmrshiftdb2 is the only open‐source 13NMR chemical shift predictor that can be freely operated in batch mode. CFM‐ID 4.0 is one of the best‐performing open‐source tools for ESI‐MS/MS spectra prediction. LOTUS is a freely usable and comprehensive collection of secondary metabolites. Integrating the open source database and software LOTUS, CFM‐ID, and nmrshiftdb2 in a dereplication workflow requires presently programming skills, owing to the diversity of data encoding and processing procedures. A graphical user interface that integrates seamlessly chemical structure collection, spectral data prediction and database building still does not exist, as far as we know. The present work proposes a stand–alone software tool that assists the identification of mixture components in a simple way.

Controlling Swing Rates in Macrocyclic Molecular Mortise Hinges.

Hinge motion is observed in macrocyclic, mortise-type molecular hinges using variable temperature NMR spectroscopy. The data is consistent with dynamic hinging from a folded-to-extended-to-folded enantiomeric state. Crystallographic and solution structures of the folded states are reported. Chemical shift predictions derived from crystallographic data corroborate fully revolute hinge motion. The rate of hinging is affected by steric congestion at the hinge axis. A macrocycle containing glycine, 1, hinges faster than one comprising aminoisobutyric acid, 2. The free energies of activation, ΔG≠ , for 1 and 2 were determined to be 13.3±0.3 kcal/mol and 16.3±0.3 kcal/mol, respectively. This barrier is largely independent of solvent across those surveyed (CD3 OD, CD3 CN, DMSO-d6 , pyridine-d5 , D2 O). Experiment and computation predict energy barriers that are consistent with disruption of an intramolecular network of hydrogen bonds. DFT calculations reveal a pathway for hinge motion.

Alpiniamides E-G from the Saline Lake-Derived Streptomyces sp. QHA48 and Their Lipid Accumulation Inhibitory Activity.

Alpiniamides E-G, three previously unreported linear polyketide derivatives, along with two known compounds, were isolated from Streptomyces sp. QHA48, which was isolated from the saline lakes of Qinghai-Tibet Plateau. The structures of these compounds were determined through analysis of their spectroscopic data, as well as density functional theory prediction of NMR chemical shifts, application of the DP4+ algorithm and electronic circular dichroism (ECD) calculations. In a cell-based lipid-lowering assay, all five alpiniamides exhibited significant inhibition of lipid accumulation in HepG2 cells without inducing cytotoxic effects at a concentration of 27 μM.

(-)-Amarbellisine is none other than (-)-montanine: A CASE-3D and dereplication NMR study

The 2013 total synthesis of the putative structure of the (-)-amarbellisine lycorine alkaloid reported very significant differences between the 1H and 13C NMR spectra of the synthesized product and the originally isolated alkaloid, and it was hypothesized that the correct structure should be a diastereoisomer of the originally proposed one. The present CASE-3D configurational analysis, based on DFT predictions of 13C chemical shifts, provided very bad fits for all possible diastereoisomeric lycorine-type structures and reexamination of the NMR data reported for natural (-)-amarbellisine revealed several inconsistencies with the proposed lycorine-like structure. Dereplication of the original 13C shifts and literature searching showed that (-)-amarbellisine is none other than the well-known (-)-montanine alkaloid.

Do Models beyond Hybrid Density Functionals Increase the Agreement with Experiment for Predicted NMR Chemical Shifts or Electric Field Gradient Tensors in Organic Solids?

Ab initio predictions of chemical shifts and electric field gradient (EFG) tensor components are frequently used to help interpret solid-state nuclear magnetic resonance (NMR) experiments. Typically, these predictions employ density functional theory (DFT) with generalized gradient approximation (GGA) functionals, though hybrid functionals have been shown to improve accuracy relative to experiment. Here, the performance of a dozen models beyond the GGA approximation are examined for the prediction of solid-state NMR observables, including meta-GGA, hybrid, and double-hybrid density functionals and second-order Møller-Plesset perturbation theory (MP2). These models are tested on organic molecular crystal data sets containing 169 experimental 13C and 15N chemical shifts and 114 17O and 14N EFG tensor components. To make these calculations affordable, gauge-including projector augmented wave (GIPAW) Perdew-Burke-Ernzerhof (PBE) calculations with periodic boundary conditions are combined with a local intramolecular correction computed at the higher level of theory. Within the context of typical NMR property calculations performed on a static, DFT-optimized crystal structure, the benchmarking finds that the double-hybrid DFT functionals produce errors versus experiment that are no smaller than those of hybrid functionals in the best cases, and they can be larger. MP2 errors versus experiment are even bigger. Overall, no practical advantages are found for using any of the tested double-hybrid functionals or MP2 to predict experimental solid-state NMR chemical shifts and EFG tensor components for routine organic crystals, especially given the higher computational cost of those methods. This finding likely reflects error cancellation benefiting the hybrid functionals. Improving the accuracy of the predicted chemical shifts and EFG tensors relative to experiment would probably require more robust treatments of the crystal structures, their dynamics, and other factors.

Dendrillic Acids A and B: Nitrogenous, Rearranged Spongian Nor-Diterpenes from a Dendrilla sp. Marine Sponge.

Two nitrogenous rearranged spongian nor-diterpenoids, dendrillic acids A and B, were isolated from a marine sponge Dendrilla sp. (order: Dendroceratida; family: Darwinellidae). The structures of the metabolites were elucidated on the basis of spectroscopic analysis as well as density functional theory prediction of NMR chemical shifts and application of the DP4+ algorithm. The absolute configuration of the metabolites was established via comparison of experimental and time-dependent density functional theory predicted electronic circular dichroism data. An unusual epimerization reaction was observed leading to the interconversion of the metabolites upon storage in dimethyl sulfoxide solution, which is proposed to proceed via an anionic pathway as probed via isotopic incorporation experiments. Evaluation against a panel of micro-organisms and cell lines revealed that the compounds were devoid of any significant biological activity against all organisms tested, with the exception of mild antiprotozoal activity displayed by dendrillic acid B (2) against Giardia duodenalis.