A deep generative model enables automated structure elucidation of novel psychoactive substances

Knowledge-based in silico fragmentation and annotation of mass spectra for natural products with MassKG

Liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) is a potent analytical technique utilized for identifying natural products from complex sources. However, due to the structural diversity, annotating LC-MS/MS data of natural products efficiently remains challenging, hindering the discovery process of novel active structures. Here, we introduce MassKG, an algorithm that combines a knowledge-based fragmentation strategy and a deep learning-based molecule generation model to aid in rapid dereplication and the discovery of novel NP structures. Specifically, MassKG has compiled 407,720 known NP structures and, based on this, generated 266,353 new structures using chemical language models for the discovery of potential novel compounds. Furthermore, MassKG demonstrates exceptional performance in spectra annotation compared to state-of-the-art algorithms. To enhance usability, MassKG has been implemented as a web server for annotating tandem mass spectral data (MS/MS, MS2) with a user-friendly interface, automatic reporting, and fragment tree visualization. Lastly, the interpretive capability of MassKG is comprehensively validated through composition analysis and MS annotation of Panax notoginseng, Ginkgo biloba, Codonopsis pilosula, and Astragalus membranaceus. MassKG is now accessible at https://xomics.com.cn/masskg.

Artificial intelligence with mass spectrometry-based multimodal molecular profiling methods for advancing therapeutic discovery of infectious diseases

Infectious diseases, driven by a diverse array of pathogens, can swiftly undermine public health systems. Accurate diagnosis and treatment of infectious diseases-centered around the identification of biomarkers and the elucidation of disease mechanisms-are in dire need of more versatile and practical analytical approaches. Mass spectrometry (MS)-based molecular profiling methods can deliver a wealth of information on a range of functional molecules, including nucleic acids, proteins, and metabolites. While MS-driven omics analyses can yield vast datasets, the sheer complexity and multi-dimensionality of MS data can significantly hinder the identification and characterization of functional molecules within specific biological processes and events. Artificial intelligence (AI) emerges as a potent complementary tool that can substantially enhance the processing and interpretation of MS data. AI applications in this context lead to the reduction of spurious signals, the improvement of precision, the creation of standardized analytical frameworks, and the increase of data integration efficiency. This critical review emphasizes the pivotal roles of MS based omics strategies in the discovery of biomarkers and the clarification of infectious diseases. Additionally, the review underscores the transformative ability of AI techniques to enhance the utility of MS-based molecular profiling in the field of infectious diseases by refining the quality and practicality of data produced from omics analyses. In conclusion, we advocate for a forward-looking strategy that integrates AI with MS-based molecular profiling. This integration aims to transform the analytical landscape and the performance of biological molecule characterization, potentially down to the single-cell level. Such advancements are anticipated to propel the development of AI-driven predictive models, thus improving the monitoring of diagnostics and therapeutic discovery for the ongoing challenge related to infectious diseases.

[The challenges for psychiatric care posed by synthetic drugs]

BACKGROUND

In addition to the drugs that have been known for decades, several hundred mainly synthetic substances have been identified as drugs for the first time in the last 20 years.

AIM OF THE WORK

Presentation of the various groups of substances and their psychotropic effects, the epidemiology of their use and the legal and social background of this development.

MATERIAL

Narrative literature review.

RESULTS

The most important new psychoactive substances (NPS) are synthetic cannabinoids, synthetic stimulants (cathinones), halluginogens and new synthetic opioids (NSO), in particular fentanyl and related substances. The new substances do not have any qualitatively new psychotropic effects. They were brought onto the market in particular as substitutes for substances subject to the Narcotics Act but are often associated with dangerous side effects and even mortality. The increasing availability of these substances has gone hand in hand with the establishment of the Internet as a source of knowledge (e.g. for synthesis routes) and as a marketplace. Substance group-related regulations have also been established in Germany (New Psychoactive Substances Act). In Germany the prevalence of NPS use is significantly lower than that of cannabis; however, there are indications that the production and distribution of synthetic drugs is more profitable for drug dealers than with conventional plant-based drugs, such as heroin. In the USA, for example, NSOs are the primarily drugs used for opioid addiction.

DISCUSSION

It remains to be seen whether NPS and NSOs will replace conventional drugs. The availability of synthetic drugs is more difficult to reduce than that of plant-based drugs. Harm reduction measures should be expanded, e.g., early warning systems for new drugs, drug checking and naloxone programs.

Identification of synthetic cathinone positional isomers using electron activated dissociation mass spectrometry

BACKGROUND

Synthetic cathinones (SCs) are a large category of new psychoactive substances (NPS), which pose a serious threat to public health due to limited information about their toxicology and pharmacology. Many SCs are closely related in their chemical structures, with some substances being positional isomers. In this study, we propose a new workflow for the identification of SC isomers using liquid chromatography-high-resolution tandem mass spectrometry (LC-HRMS2) combined with electron activated dissociation (EAD) and chemometrics. Differentiation between isomeric SCs is essential for both legislative and public safety reasons, since minor differences in their molecular structures may change their legal status and pharmacological profiles.

RESULTS

The workflow was optimized using ring-substituted isomers of methylmethcathinones, methylethcathinones, and chloromethcathinones. The kinetic energy in the EAD cell was investigated at three levels (i.e., 15, 18, and 20 eV) for each group. Two data analysis methods (i.e., t-distributed stochastic neighbor embedding [t-SNE] and a Random Forest [RF] algorithm) were applied using the obtained EAD mass spectral data. The three sets of ring-substituted SCs were clearly distinguished using t-SNE and an RF algorithm. Moreover, the RF approach resulted in a 97 % classification accuracy for isomer identification using various combinations of compounds, isomers, and electron kinetic energies. This workflow was subsequentially applied to the analysis of 26 blind street samples, resulting in a 92 % classification accuracy for isomer identification. However, the accuracy varied based on the kinetic electron energy. A subset of the original data set, focusing on 15-eV data only, was used, resulting in a classification accuracy of 100 %.

SIGNIFICANCE

This study presents the first LC-HRMS2 workflow based on EAD and chemometrics, which resulted in a classification accuracy of 100 % of authentic street samples. The developed LC-HRMS2 workflow demonstrates that EAD product ions and their characteristic ion ratios can be successfully used to identify ring-substituted positional isomers of SCs.

Critical review on in silico methods for structural annotation of chemicals detected with LC/HRMS non-targeted screening

Non-targeted screening with liquid chromatography coupled to high-resolution mass spectrometry (LC/HRMS) is increasingly leveraging in silico methods, including machine learning, to obtain candidate structures for structural annotation of LC/HRMS features and their further prioritization. Candidate structures are commonly retrieved based on the tandem mass spectral information either from spectral or structural databases; however, the vast majority of the detected LC/HRMS features remain unannotated, constituting what we refer to as a part of the unknown chemical space. Recently, the exploration of this chemical space has become accessible through generative models. Furthermore, the evaluation of the candidate structures benefits from the complementary empirical analytical information such as retention time, collision cross section values, and ionization type. In this critical review, we provide an overview of the current approaches for retrieving and prioritizing candidate structures. These approaches come with their own set of advantages and limitations, as we showcase in the example of structural annotation of ten known and ten unknown LC/HRMS features. We emphasize that these limitations stem from both experimental and computational considerations. Finally, we highlight three key considerations for the future development of in silico methods.

Synthetic biology advances towards a bio-based society in the era of artificial intelligence

Synthetic biology is a rapidly emerging field with broad underlying applications in health, industry, agriculture, or environment, enabling sustainable solutions for unmet needs of modern society. With the very recent addition of artificial intelligence (AI) approaches, this field is now growing at a rate that can help reach the envisioned goals of bio-based society within the next few decades. Integrating AI with plant-based technologies, such as protein engineering, phytochemicals production, plant system engineering, or microbiome engineering, potentially disruptive applications have already been reported. These include enzymatic synthesis of new-to-nature molecules, bioelectricity generation, or biomass applications as construction material. Thus, in the not-so-distant future, synthetic biologists will help attain the overarching goal of a sustainable yet efficient production system for every aspect of society.

Deep Learning-Assisted Spectrum-Structure Correlation: State-of-the-Art and Perspectives

Spectrum-structure correlation is playing an increasingly crucial role in spectral analysis and has undergone significant development in recent decades. With the advancement of spectrometers, the high-throughput detection triggers the explosive growth of spectral data, and the research extension from small molecules to biomolecules accompanies massive chemical space. Facing the evolving landscape of spectrum-structure correlation, conventional chemometrics becomes ill-equipped, and deep learning assisted chemometrics rapidly emerges as a flourishing approach with superior ability of extracting latent features and making precise predictions. In this review, the molecular and spectral representations and fundamental knowledge of deep learning are first introduced. We then summarize the development of how deep learning assist to establish the correlation between spectrum and molecular structure in the recent 5 years, by empowering spectral prediction (i.e., forward structure-spectrum correlation) and further enabling library matching and de novo molecular generation (i.e., inverse spectrum-structure correlation). Finally, we highlight the most important open issues persisted with corresponding potential solutions. With the fast development of deep learning, it is expected to see ultimate solution of establishing spectrum-structure correlation soon, which would trigger substantial development of various disciplines.

Structural annotation of unknown molecules in a miniaturized mass spectrometer based on a transformer enabled fragment tree method

Structural annotation of small molecules in tandem mass spectrometry has always been a central challenge in mass spectrometry analysis, especially using a miniaturized mass spectrometer for on-site testing. Here, we propose the Transformer enabled Fragment Tree (TeFT) method, which combines various types of fragmentation tree models and a deep learning Transformer module. It is aimed to generate the specific structure of molecules de novo solely from mass spectrometry spectra. The evaluation results on different open-source databases indicated that the proposed model achieved remarkable results in that the majority of molecular structures of compounds in the test can be successfully recognized. Also, the TeFT has been validated on a miniaturized mass spectrometer with low-resolution spectra for 16 flavonoid alcohols, achieving complete structure prediction for 8 substances. Finally, TeFT confirmed the structure of the compound contained in a Chinese medicine substance called the Anweiyang capsule. These results indicate that the TeFT method is suitable for annotating fragmentation peaks with clear fragmentation rules, particularly when applied to on-site mass spectrometry with lower mass resolution.

Cross-Modal Retrieval Between 13C NMR Spectra and Structures Based on Focused Libraries

Library matching by comparing carbon-13 nuclear magnetic resonance (13C NMR) spectra with spectral data in the library is a crucial method for compound identification. In our previous paper, we introduced a deep contrastive learning system called CReSS, which used a library that contained more structures. However, CReSS has two limitations: there were no unknown structures in the library, and a redundant library reduces the structure-elucidation accuracy. Herein, we replaced the oversize traditional libraries with focused libraries containing a small number of molecules. A previously generative model, CMGNet, was used to generate focused libraries for CReSS. The combined model achieved a Top-10 accuracy of 54.03% when tested on 6,471 13C NMR spectra. In comparison, CReSS with a random reference structure library achieved an accuracy of only 9.17%. Furthermore, to expand the advantages of the focused libraries, we proposed SAmpRNN, which is a recurrent neural network (RNN). With the large focused library amplified by SAmpRNN, the structure-identification accuracy of the model increased in 70.0% of the 30 random example cases. In general, cross-modal retrieval between 13C NMR spectra and structures based on focused libraries (CFLS) achieved high accuracy and provided more accurate candidate structures than traditional libraries for compound identification.

AI is a viable alternative to high throughput screening: a 318-target study

High throughput screening (HTS) is routinely used to identify bioactive small molecules. This requires physical compounds, which limits coverage of accessible chemical space. Computational approaches combined with vast on-demand chemical libraries can access far greater chemical space, provided that the predictive accuracy is sufficient to identify useful molecules. Through the largest and most diverse virtual HTS campaign reported to date, comprising 318 individual projects, we demonstrate that our AtomNet® convolutional neural network successfully finds novel hits across every major therapeutic area and protein class. We address historical limitations of computational screening by demonstrating success for target proteins without known binders, high-quality X-ray crystal structures, or manual cherry-picking of compounds. We show that the molecules selected by the AtomNet® model are novel drug-like scaffolds rather than minor modifications to known bioactive compounds. Our empirical results suggest that computational methods can substantially replace HTS as the first step of small-molecule drug discovery.

PS2MS: A Deep Learning-Based Prediction System for Identifying New Psychoactive Substances Using Mass Spectrometry

The rapid proliferation of new psychoactive substances (NPS) poses significant challenges to conventional mass-spectrometry-based identification methods due to the absence of reference spectra for these emerging substances. This paper introduces PS2MS, an AI-powered predictive system designed specifically to address the limitations of identifying the emergence of unidentified novel illicit drugs. PS2MS builds a synthetic NPS database by enumerating feasible derivatives of known substances and uses deep learning to generate mass spectra and chemical fingerprints. When the mass spectrum of an analyte does not match any known reference, PS2MS simultaneously examines the chemical fingerprint and mass spectrum against the putative NPS database using integrated metrics to deduce possible identities. Experimental results affirm the effectiveness of PS2MS in identifying cathinone derivatives within real evidence specimens, signifying its potential for practical use in identifying emerging drugs of abuse for researchers and forensic experts.

Unlocking the Potential of High-Quality Dopamine Transporter Pharmacological Data: Advancing Robust Machine Learning-Based QSAR Modeling

The dopamine transporter (DAT) plays a critical role in the central nervous system and has been implicated in numerous psychiatric disorders. The ligand-based approaches are instrumental to decipher the structure-activity relationship (SAR) of DAT ligands, especially the quantitative SAR (QSAR) modeling. By gathering and analyzing data from literature and databases, we systematically assemble a diverse range of ligands binding to DAT, aiming to discern the general features of DAT ligands and uncover the chemical space for potential novel DAT ligand scaffolds. The aggregation of DAT pharmacological activity data, particularly from databases like ChEMBL, provides a foundation for constructing robust QSAR models. The compilation and meticulous filtering of these data, establishing high-quality training datasets with specific divisions of pharmacological assays and data types, along with the application of QSAR modeling, prove to be a promising strategy for navigating the pertinent chemical space. Through a systematic comparison of DAT QSAR models using training datasets from various ChEMBL releases, we underscore the positive impact of enhanced data set quality and increased data set size on the predictive power of DAT QSAR models.

Behind every good research there are data. What are they and their importance to forensic science

Data underpinning science have become one of the most precious assets in research, and while the principles of FAIR (Findable, Accessible, Interoperable and Reusable) have been put forward as a guide to how to approach data handling, data sharing and long-term storage still remain a challenge for many research areas including forensic science. The reporting and the sharing of data can be made easier by giving them structure, the use of suitable labels and the inclusion of descriptors collated into metadata prior to their deposition in repositories with persistent identifiers. Such a systematic approach would strengthen the quality and the integrity of research while providing greater transparency to published materials.

Deep Learning-Enabled MS/MS Spectrum Prediction Facilitates Automated Identification Of Novel Psychoactive Substances

The market for illicit drugs has been reshaped by the emergence of more than 1100 new psychoactive substances (NPS) over the past decade, posing a major challenge to the forensic and toxicological laboratories tasked with detecting and identifying them. Tandem mass spectrometry (MS/MS) is the primary method used to screen for NPS within seized materials or biological samples. The most contemporary workflows necessitate labor-intensive and expensive MS/MS reference standards, which may not be available for recently emerged NPS on the illicit market. Here, we present NPS-MS, a deep learning method capable of accurately predicting the MS/MS spectra of known and hypothesized NPS from their chemical structures alone. NPS-MS is trained by transfer learning from a generic MS/MS prediction model on a large data set of MS/MS spectra. We show that this approach enables a more accurate identification of NPS from experimentally acquired MS/MS spectra than any existing method. We demonstrate the application of NPS-MS to identify a novel derivative of phencyclidine (PCP) within an unknown powder seized in Denmark without the use of any reference standards. We anticipate that NPS-MS will allow forensic laboratories to identify more rapidly both known and newly emerging NPS. NPS-MS is available as a web server at https://nps-ms.ca/, which provides MS/MS spectra prediction capabilities for given NPS compounds. Additionally, it offers MS/MS spectra identification against a vast database comprising approximately 8.7 million predicted NPS compounds from DarkNPS and 24.5 million predicted ESI-QToF-MS/MS spectra for these compounds.

Identification of Emerging Novel Psychoactive Substances by Retrospective Analysis of Population-Scale Mass Spectrometry Data Sets

Over the last two decades, hundreds of new psychoactive substances (NPSs), also known as "designer drugs", have emerged on the illicit drug market. The toxic and potentially fatal effects of these compounds oblige laboratories around the world to screen for NPS in seized materials and biological samples, commonly using high-resolution mass spectrometry. However, unambiguous identification of a NPS by mass spectrometry requires comparison to data from analytical reference materials, acquired on the same instrument. The sheer number of NPSs that are available on the illicit market, and the pace at which new compounds are introduced, means that forensic laboratories must make difficult decisions about which reference materials to acquire. Here, we asked whether retrospective suspect screening of population-scale mass spectrometry data could provide a data-driven platform to prioritize emerging NPSs for assay development. We curated a suspect database of precursor and diagnostic fragment ion masses for 83 emerging NPSs and used this database to retrospectively screen mass spectrometry data from 12,727 urine drug screens from one Canadian province. We developed integrative computational strategies to prioritize the most reliable identifications and tracked the frequency of these identifications over a 3 year study period between August 2019 and August 2022. The resulting data were used to guide the acquisition of new reference materials, which were in turn used to validate a subset of the retrospective identifications. Last, we took advantage of matching clinical reports for all 12,727 samples to systematically benchmark the accuracy of our retrospective data analysis approach. Our work opens up new avenues to enable the rapid detection of emerging illicit drugs through large-scale reanalysis of mass spectrometry data.

Hallucinating hallucinogens

Fighting the designer drug epidemic with generative AI.

Advances in the Application of Artificial Intelligence-Based Spectral Data Interpretation: A Perspective

The interpretation of spectral data, including mass, nuclear magnetic resonance, infrared, and ultraviolet-visible spectra, is critical for obtaining molecular structural information. The development of advanced sensing technology has multiplied the amount of available spectral data. Chemical experts must use basic principles corresponding to the spectral information generated by molecular fragments and functional groups. This is a time-consuming process that requires a solid professional knowledge base. In recent years, the rapid development of computer science and its applications in cheminformatics and the emergence of computer-aided expert systems have greatly reduced the difficulty in analyzing large quantities of data. For expert systems, however, the problem-solving strategy must be known in advance or extracted by human experts and translated into algorithms. Gratifyingly, the development of artificial intelligence (AI) methods has shown great promise for solving such problems. Traditional algorithms, including the latest neural network algorithms, have shown great potential for both extracting useful information and processing massive quantities of data. This Perspective highlights recent innovations covering all of the emerging AI-based spectral interpretation techniques. In addition, the main limitations and current obstacles are presented, and the corresponding directions for further research are proposed. Moreover, this Perspective gives the authors' personal outlook on the development and future applications of spectral interpretation.

67 million natural product-like compound database generated via molecular language processing

Natural products are a rich resource of bioactive compounds for valuable applications across multiple fields such as food, agriculture, and medicine. For natural product discovery, high throughput in silico screening offers a cost-effective alternative to traditional resource-heavy assay-guided exploration of structurally novel chemical space. In this data descriptor, we report a characterized database of 67,064,204 natural product-like molecules generated using a recurrent neural network trained on known natural products, demonstrating a significant 165-fold expansion in library size over the approximately 400,000 known natural products. This study highlights the potential of using deep generative models to explore novel natural product chemical space for high throughput in silico discovery.

Data-Driven Elucidation of Flavor Chemistry

Flavor molecules are commonly used in the food industry to enhance product quality and consumer experiences but are associated with potential human health risks, highlighting the need for safer alternatives. To address these health-associated challenges and promote reasonable application, several databases for flavor molecules have been constructed. However, no existing studies have comprehensively summarized these data resources according to quality, focused fields, and potential gaps. Here, we systematically summarized 25 flavor molecule databases published within the last 20 years and revealed that data inaccessibility, untimely updates, and nonstandard flavor descriptions are the main limitations of current studies. We examined the development of computational approaches (e.g., machine learning and molecular simulation) for the identification of novel flavor molecules and discussed their major challenges regarding throughput, model interpretability, and the lack of gold-standard data sets for equitable model evaluation. Additionally, we discussed future strategies for the mining and designing of novel flavor molecules based on multi-omics and artificial intelligence to provide a new foundation for flavor science research.

Molecular Generative Model via Retrosynthetically Prepared Chemical Building Block Assembly

Deep generative models are attracting attention as a smart molecular design strategy. However, previous models often render molecules with low synthesizability, hindering their real-world applications. Here, a novel graph-based conditional generative model which makes molecules by tailoring retrosynthetically prepared chemical building blocks until achieving target properties in an auto-regressive fashion is proposed. This strategy improves the synthesizability and property control of the resulting molecules and also helps learn how to select appropriate building blocks and bind them together to achieve target properties. By applying a negative sampling method to the selection process of building blocks, this model overcame a critical limitation of previous fragment-based models, which can only use molecules from the training set during generation. As a result, the model works equally well with unseen building blocks without sacrificing computational efficiency. It is demonstrated that the model can generate potential inhibitors with high docking scores against the 3CL protease of SARS-COV-2.

Machine Learning-Assisted Rapid Screening of Four Types of New Psychoactive Substances in Drug Seizures

Over the past few years, new psychoactive substances (NPS) have become a global health and social problem because of their wide variety, constant structural renewal, vague legal definitions, and rapid adaptation to legal restrictions. The rapid structural modifications of NPS have posed significant challenges for the screening and identification of these new substances using traditional mass spectrometric techniques based on reference substances or a mass spectral database. Here, we propose supervised machine learning (ML) classification models such as k-nearest neighbors, support vector machine, random forest, and multigrained cascade forest for the rapid screening of NPS using mass spectrometric data. This approach utilizes ML methods to learn the statistical probability distributions of mass spectral data for NPS and non-NPS. Four classification ML models were generated and evaluated using a data set comprising 567 LC-MS and 732 GC-MS spectra. Through cross validation, we achieved an F1 score of 0.35-0.97. These algorithms were applied in conjunction with mass spectrometry techniques for the detection of six seizures including electronic cigarette oil and suspected powdered substances netted in drug trafficking cases. The models provided warning signals for synthetic cannabinoids, synthetic cathinones, and fentanyl. Thus, an early warning system was successfully established, which provided a useful method for reliable and effective identifications of unknown NPS.

A smile is all you need: predicting limiting activity coefficients from SMILES with natural language processing

The knowledge of mixtures' phase equilibria is crucial in nature and technical chemistry. Phase equilibria calculations of mixtures require activity coefficients. However, experimental data on activity coefficients are often limited due to the high cost of experiments. For an accurate and efficient prediction of activity coefficients, machine learning approaches have been recently developed. However, current machine learning approaches still extrapolate poorly for activity coefficients of unknown molecules. In this work, we introduce a SMILES-to-properties-transformer (SPT), a natural language processing network, to predict binary limiting activity coefficients from SMILES codes. To overcome the limitations of available experimental data, we initially train our network on a large dataset of synthetic data sampled from COSMO-RS (10 million data points) and then fine-tune the model on experimental data (20 870 data points). This training strategy enables the SPT to accurately predict limiting activity coefficients even for unknown molecules, cutting the mean prediction error in half compared to state-of-the-art models for activity coefficient predictions such as COSMO-RS and UNIFACDortmund, and improving on recent machine learning approaches.

Predicting the structure of unexplored novel fentanyl analogues by deep learning model

Fentanyl and its analogues are psychoactive substances and the concern of fentanyl abuse has been existed in decades. Because the structure of fentanyl is easy to be modified, criminals may synthesize new fentanyl analogues to avoid supervision. The drug supervision is based on the structure matching to the database and too few kinds of fentanyl analogues are included in the database, so it is necessary to find out more potential fentanyl analogues and expand the sample space of fentanyl analogues. In this study, we introduced two deep generative models (SeqGAN and MolGPT) to generate potential fentanyl analogues, and a total of 11 041 valid molecules were obtained. The results showed that not only can we generate molecules with similar property distribution of original data, but the generated molecules also contain potential fentanyl analogues that are not pretty similar to any of original data. Ten molecules based on the rules of fentanyl analogues were selected for NMR, MS and IR validation. The results indicated that these molecules are all unreported fentanyl analogues. Furthermore, this study is the first to apply the deep learning to the generation of fentanyl analogues, greatly expands the exploring space of fentanyl analogues and provides help for the supervision of fentanyl.

MSNovelist: de novo structure generation from mass spectra

Current methods for structure elucidation of small molecules rely on finding similarity with spectra of known compounds, but do not predict structures de novo for unknown compound classes. We present MSNovelist, which combines fingerprint prediction with an encoder–decoder neural network to generate structures de novo solely from tandem mass spectrometry (MS²) spectra. In an evaluation with 3,863 MS² spectra from the Global Natural Product Social Molecular Networking site, MSNovelist predicted 25% of structures correctly on first rank, retrieved 45% of structures overall and reproduced 61% of correct database annotations, without having ever seen the structure in the training phase. Similarly, for the CASMI 2016 challenge, MSNovelist correctly predicted 26% and retrieved 57% of structures, recovering 64% of correct database annotations. Finally, we illustrate the application of MSNovelist in a bryophyte MS² dataset, in which de novo structure prediction substantially outscored the best database candidate for seven spectra. MSNovelist is ideally suited to complement library-based annotation in the case of poorly represented analyte classes and novel compounds.

MSNovelist combines fingerprint prediction with an encoder–decoder neural network for de novo structure generation of small molecules from mass spectra.

Developments in high-resolution mass spectrometric analyses of new psychoactive substances

The proliferation of new psychoactive substances (NPS) has necessitated the development and improvement of current practices for the detection and identification of known NPS and newly emerging derivatives. High-resolution mass spectrometry (HRMS) is quickly becoming the industry standard for these analyses due to its ability to be operated in data-independent acquisition (DIA) modes, allowing for the collection of large amounts of data and enabling retrospective data interrogation as new information becomes available. The increasing popularity of HRMS has also prompted the exploration of new ways to screen for NPS, including broad-spectrum wastewater analysis to identify usage trends in the community and metabolomic-based approaches to examine the effects of drugs of abuse on endogenous compounds. In this paper, the novel applications of HRMS techniques to the analysis of NPS is reviewed. In particular, the development of innovative data analysis and interpretation approaches is discussed, including the application of machine learning and molecular networking to toxicological analyses.

MassGenie: A Transformer-Based Deep Learning Method for Identifying Small Molecules from Their Mass Spectra

The 'inverse problem' of mass spectrometric molecular identification ('given a mass spectrum, calculate/predict the 2D structure of the molecule whence it came') is largely unsolved, and is especially acute in metabolomics where many small molecules remain unidentified. This is largely because the number of experimentally available electrospray mass spectra of small molecules is quite limited. However, the forward problem ('calculate a small molecule's likely fragmentation and hence at least some of its mass spectrum from its structure alone') is much more tractable, because the strengths of different chemical bonds are roughly known. This kind of molecular identification problem may be cast as a language translation problem in which the source language is a list of high-resolution mass spectral peaks and the 'translation' a representation (for instance in SMILES) of the molecule. It is thus suitable for attack using the deep neural networks known as transformers. We here present MassGenie, a method that uses a transformer-based deep neural network, trained on ~6 million chemical structures with augmented SMILES encoding and their paired molecular fragments as generated in silico, explicitly including the protonated molecular ion. This architecture (containing some 400 million elements) is used to predict the structure of a molecule from the various fragments that may be expected to be observed when some of its bonds are broken. Despite being given essentially no detailed nor explicit rules about molecular fragmentation methods, isotope patterns, rearrangements, neutral losses, and the like, MassGenie learns the effective properties of the mass spectral fragment and valency space, and can generate candidate molecular structures that are very close or identical to those of the 'true' molecules. We also use VAE-Sim, a previously published variational autoencoder, to generate candidate molecules that are 'similar' to the top hit. In addition to using the 'top hits' directly, we can produce a rank order of these by 'round-tripping' candidate molecules and comparing them with the true molecules, where known. As a proof of principle, we confine ourselves to positive electrospray mass spectra from molecules with a molecular mass of 500Da or lower, including those in the last CASMI challenge (for which the results are known), getting 49/93 (53%) precisely correct. The transformer method, applied here for the first time to mass spectral interpretation, works extremely effectively both for mass spectra generated in silico and on experimentally obtained mass spectra from pure compounds. It seems to act as a Las Vegas algorithm, in that it either gives the correct answer or simply states that it cannot find one. The ability to create and to 'learn' millions of fragmentation patterns in silico, and therefrom generate candidate structures (that do not have to be in existing libraries) directly, thus opens up entirely the field of de novo small molecule structure prediction from experimental mass spectra.