Deep Learning-Based Method for Compound Identification in NMR Spectra of Mixtures

Wei, Weiwei; Liao, Yuxuan; Wang, Yufei; Wang, Shaoqi; Du, Wen; Lü, Hongmei; Kong, Bo; Yang, Huawu; Zhang, Zhimin

doi:10.3390/molecules27123653

Cited by 24 publications

(21 citation statements)

References 79 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We built a spectral library with the NMR spectra of 74 small molecules. Both experimental and predicted NMR spectra were used in the library based on data availability from a peer-reviewed publication [ 40 ] and the Human Metabolome Database [ 51 ]. The corresponding WPT spectral library for the molecules was computed using the Daubechies-9 (Db9) wavelet, and a full reduction of all multiplets to singlets in a spectrum was used as the criterion to select the optimal decomposition level.…”

Section: Methodsmentioning

confidence: 99%

“…One of its major drawbacks is the requirement of prior knowledge about the scalar coupling patterns and the coupling constants, which is reasonable for

C NMR, but unsuitable for

H NMR spectroscopy. Apart from this, a series of spectral analysis tools has been developed, which include peak matching strategies [ 33 , 34 , 35 ], spectral editing [ 36 , 37 ], similarity measure [ 38 , 39 ], and deep-learning-based tools [ 40 , 41 , 42 ], for identifying small molecule mixture constituents from the corresponding NMR spectra. However, those applications can be seldom generalized, often suffer from low reliability, and/or require extremely large and specifically designed training datasets.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised Analysis of Small Molecule Mixtures by Wavelet-Based Super-Resolved NMR

Roy

Srivastava

2023

Molecules

View full text Add to dashboard Cite

Resolving small molecule mixtures by nuclear magnetic resonance (NMR) spectroscopy has been of great interest for a long time for its precision, reproducibility, and efficiency. However, spectral analyses for such mixtures are often highly challenging due to overlapping resonance lines and limited chemical shift windows. The existing experimental and theoretical methods to produce shift NMR spectra in dealing with the problem have limited applicability owing to sensitivity issues, inconsistency, and/or the requirement of prior knowledge. Recently, we resolved the problem by decoupling multiplet structures in NMR spectra by the wavelet packet transform (WPT) technique. In this work, we developed a scheme for deploying the method in generating highly resolved WPT NMR spectra and predicting the composition of the corresponding molecular mixtures from their 1H NMR spectra in an automated fashion. The four-step spectral analysis scheme consists of calculating the WPT spectrum, peak matching with a WPT shift NMR library, followed by two optimization steps in producing the predicted molecular composition of a mixture. The robustness of the method was tested on an augmented dataset of 1000 molecular mixtures, each containing 3 to 7 molecules. The method successfully predicted the constituent molecules with a median true positive rate of 1.0 against the varying compositions, while a median false positive rate of 0.04 was obtained. The approach can be scaled easily for much larger datasets.

show abstract

Section: Methodsmentioning

confidence: 99%

“…One of its major drawbacks is the requirement of prior knowledge about the scalar coupling patterns and the coupling constants, which is reasonable for

C NMR, but unsuitable for

Section: Introductionmentioning

confidence: 99%

Unsupervised Analysis of Small Molecule Mixtures by Wavelet-Based Super-Resolved NMR

Roy

Srivastava

2023

Molecules

View full text Add to dashboard Cite

show abstract

“…We built a spectral library with the NMR spectra of 74 small molecules. Both experimental and predicted NMR spectra were used in the library based on data availability from a peer-reviewed publication [35] and the Human Metabolome Database [39]. The corresponding WPT spectral library for the molecules was computed using Daubechies-9 (Db9) wavelet and a full reduction of all multiplets to singlets in a spectrum was used as the criterion to select the optimal decomposition level.…”

Section: Spectral Library and Augmented Dataset Creationmentioning

confidence: 99%

“…2 of 11 [35][36][37], for identifying small molecule mixture constituents from the corresponding NMR spectra. However, those applications can be seldom generalized, often suffer from low reliability and / or require extremely large and specifically designed training datasets.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised Analysis of Small Molecule Mixtures by Wavelet-Based Super-Resolved NMR

Roy¹,

Srivastava²

2022

Preprint

View full text Add to dashboard Cite

Resolving small molecule mixtures by nuclear magnetic resonance (NMR) spectroscopy has been of great interest for a long time for its precision, reproducibility and efficiency. However, spectral analyses for such mixtures are often highly challenging due to overlapping resonance lines and limited chemical shift windows. The existing experimental and theoretical methods to produce shift NMR spectra in dealing with the problem have limited applicability owing to sensitivity issues, inconsistency and / or requirement of prior knowledge. Recently, we have resolved the problem by decoupling multiplet structures in NMR spectra by the wavelet packet transform (WPT) technique. In this work, we developed a scheme for deploying the method in generating highly resolved WPT NMR spectra and predicting the composition of the corresponding molecular mixtures from their 1H NMR spectra in an automated fashion. The four-step spectral analysis scheme consists of calculating WPT spectrum, peak matching with a WPT shift NMR library, followed by two optimization steps in producing the predicted molecular composition of a mixture. The robustness of the method was tested on an augmented dataset of 1000 molecular mixtures, each containing 3 to 7 molecules. The method successfully predicted the constituent molecules with a median true positive rate of 1.0 against the varying compositions, while a median false positive rate of 0.04 was obtained. The approach can be scaled easily for much larger datasets.

show abstract

“…Besides, for other methods, rDolphin 19 and AQuA 20 , requiring manual library adjustment or dedicated sample preparations to prefilter the uninterested components within the biofluid samples, need more application convenience 21 . Recently, in deep learning, many methods have also been proposed for NMR applications ranging from spectra noise processing to metabolites identification [22][23][24] . Despite all these efforts, the metabolite features accompanied by variations in the real NMR spectra are still less capturable.…”

Section: Introductionmentioning

confidence: 99%

NMRQNet: a deep learning approach for automatic identification and quantification of metabolites using Nuclear Magnetic Resonance (NMR) in human plasma samples

Wang

Maletić-Savatić

et al. 2023

Preprint

View full text Add to dashboard Cite

Nuclear Magnetic Resonance is a powerful platform that reveals the metabolomics profiles within biofluids or tissues and contributes to personalized treatments in medical practice. However, data volume and complexity hinder the exploration of NMR spectra. Besides, the lack of fast and accurate computational tools that can handle the automatic identification and quantification of essential metabolites from NMR spectra also slows the wide application of these techniques in clinical. We present NMRQNet, a deep-learning-based pipeline for automatic identification and quantification of dominant metabolite candidates within human plasma samples. The estimated relative concentrations could be further applied in statistical analysis to extract the potential biomarkers. We evaluate our method on multiple plasma samples, including species from mice to humans, curated using three anticoagulants, covering healthy and patient conditions in neurological disorder disease, greatly expanding the metabolomics analytical space in plasma. NMRQNet accurately reconstructed the original spectra and obtained significantly better quantification results than the earlier computational methods. Besides, NMRQNet also proposed relevant metabolites biomarkers that could potentially explain the risk factors associated with the condition. NMRQNet, with improved prediction performance, highlights the limitations in the existing approaches and has shown strong application potential for future metabolomics disease studies using plasma samples.

show abstract

Deep Learning-Based Method for Compound Identification in NMR Spectra of Mixtures

Cited by 24 publications

References 79 publications

Unsupervised Analysis of Small Molecule Mixtures by Wavelet-Based Super-Resolved NMR

Unsupervised Analysis of Small Molecule Mixtures by Wavelet-Based Super-Resolved NMR

Unsupervised Analysis of Small Molecule Mixtures by Wavelet-Based Super-Resolved NMR

NMRQNet: a deep learning approach for automatic identification and quantification of metabolites using Nuclear Magnetic Resonance (NMR) in human plasma samples

Contact Info

Product

Resources

About