With new datasets, community seeks ‘AlphaFold moment’ for small molecules

AbstractTerpenoids are among the most chemically diverse natural product classes in the plant kingdom and are a source of numerous nature-inspired human medicines. One specific subclass, diterpenoids, defined by a C20 skeleton, includes attractive compounds such as the anticancer drug Taxol, the sweetener steviol, and the plant hormone gibberellin. One specialized group of ent-labdane (bicyclic) diterpenoids, primarily reported from the genus Andrographis (Acanthaceae), is the source of andrographolides, a diverse array of bioactive scaffolds with anti-inflammatory, antioxidant, anticancer, antimicrobial, and antihyperglycemic properties. Using high-throughput metabolomics, we set out to examine the molecular diversity within the Acanthaceae family, alongside other taxonomic lineages reported to accumulate diterpenoids. In a combined metabolomic investigation of 31 genera spanning Acanthaceae, Lamiaceae, and Poaceae, we annotated several reported andrographolide-type labdane diterpenoids and characteristic compounds unique to Acanthaceae. Integration of chemical, geographic, and taxonomic metadata with literature-driven assessments helped paint a chemotaxonomic and phylogeographic distribution map of andrographolide-type labdane diterpenoids across 163 countries. We found several species accumulate at least seven andrographolide-type compounds and another 17 labdane diterpenoids in our analysis. We conclude that labdane diterpenoid accumulation is more widespread, with multicontinental and shared evolutionary origins, than previously recognized.

Multi-omics-driven precision medicine (MODPM) provides a multiscale, continuously learnable framework that integrates genomics, epigenomics, transcriptomics, proteomics, metabolomics, microbiome prof...

Nature Methods - Bridging complexity and accessibility in metabolomics with MetaboApps

🚨 New Preprint Alert 🚨 Congratulations 👏 🎉 to the CompMet team for presenting this preprint on boosting library matching for untargeted mass spectrometry-based metabolomics workflows 🙌 Metabolite a...

• PostMN MassQL: https://github.com/helenamrusso/massql_post_MN • Multi-step MassQL: https://github.com/wilhan-nunes/streamlit-Multi-Step-MassQL • Reference Data-Driven: https://github.com/bittremieux-lab/gnps-rdd • Drug Readout: https://github.com/wilhan-nunes/streamlit_drug_readouts • Food Readout: https://github.com/wilhan-nunes/streamlit-food-readouts • CMMC dashboard: https://github.com/wilhan-nunes/streamlit_CMMC_analysis-dashboard • FBMN-Stats: https://github.com/Functional-Metabolomics-Lab/FBMN-STATS • Chemical Proportionality: https://github.com/Functional-Metabolomics-Lab/ChemProp-Web-App • CorrOmics: https://github.com/Functional-Metabolomics-Lab/Corromics • Conjugated Metabolome Explorer: https://github.com/Philipbear/conjugated_metabolome • Reverse Metabolomics: https://github.com/wilhan-nunes/streamlit-Reverse-Metabolomics Documentation can be accessed at https://wang-bioinformatics-lab.github.io/GNPS2_Documentation/metaboapps_overview/. For developers from the metabolomics or broader community interested in contributing new MetaboApps, we provide resources and documentation. The core infrastructure is accessible through three key repositories. The Streamlit App GNPS2 template (https://github.com/Wang-Bioinformatics-Lab/Template_Streamlit_App) offers a preconfigured template with GNPS2 integration guidelines for a quick start. GNPSDataPackage (https://github.com/Wang-Bioinformatics-Lab/GNPSDataPackage) is a Python library that enables programmatic access to GNPS2 workflow results via task IDs for postprocessing, providing the tools to work directly with analysis outputs (which are mainly CSV tables, MGF spectra files or GraphML network files). For more experienced developers, we recommend the Dash template (https://github.com/Wang-Bioinformatics-Lab/Dash_Multipage_Template), which provides a framework for building multipage applications. Pull requests are evaluated by core GNPS2 developers for code quality, efficiency and infrastructure compatibility. Developers wishing to implement new MetaboApps in GNPS2 should contact the GNPS2 team to initiate the integration and deployment process. Applications are encouraged to include clear documentation and user-friendly interfaces. We thank the...

The majority of chemical signals detected in public metabolomics repositories remain structurally undefined. Large language models (LLMs) are probabilistic systems whose capacity to generate outputs beyond their training data, which can cause hallucinations, makes them also potentially suited to hypothesize structures for molecules that have never been described. We aimed to build a system that could harness this LLM generative capacity combined with domain specific tools/framework to constrain hallucination and produce validated discoveries. We developed a GNPS2 agentic AI system that interprets LC-MS/MS data by integrating spectral alignment, molecular formula inference, rule-based structural enumeration, machine learning-based spectrum prediction, and translates natural language hypotheses from domain experts into dynamically generated analytical workflows. We demonstrate the annotation of unknown drug metabolites from public data guided by chemical hypotheses. The agent predicted, and we experimentally confirmed, a phosphorylated hydroxyzine, an acetaminophen-p-coumaric acid ester, and identified two new oxidative ibuprofen-carnitine conjugates from public repositories. These results demonstrate that LLM-driven agentic reasoning, when combined with domain expertise, can indeed generate experimentally testable structural hypotheses for previously uncharacterized metabolites leveraging pan repository data.

Curated, evidence-grounded skill and software-tool collections for scientific AI agents, generated by the AgenticScienceBuilder - HolobiomicsLab/asb-skill-collections

Metabolomics software development has accelerated rapidly, yet no recent systematic analysis has quantified how the landscape is evolving across computational methods, geographies, and the research co...

Untargeted liquid chromatography- tandem mass spectrometry (LC - MS/MS) - based metabolomics is an important technology for unbiased discovery of small molecules in biomedical (e.g., drug discovery to diagnostics), animal, plant, environmental, and microbial research. Over the past decade, ion mobility has added an additional dimension to the triplet of MS1, MS2, and retention time, helping resolve co-eluting or isomeric features in an LC- MS/MS that aid in compound identification. Here, we focused on evaluating the current trapped ion mobility spectrometry (TIMS) - amenable feature-finding tools (MZmine 4.9, MS-DIAL 5.5, and MetaboScape 2025 14.0.3) for pre-processing of metabolomics-scale data generated using a popular ion mobility mass spectrometry (IM- MS) technique, TIMS. We leveraged ten public and three benchmark TIMS datasets to evaluate these tools for their strengths and weaknesses. Our results show that MZmine consistently identified the highest number of features and confidently annotated features; however, this performance was accompanied by an increased number of false positives, due to peak splitting, as well as reduced accuracy in collision cross section (CCS) measurements. In contrast, MetaboScape achieved the highest fraction of high-quality MS2 spectra, reflecting a more conservative feature detection strategy. MS-DIAL demonstrated balanced performance, identifying features that other tools missed. Finally, we publicly release the ground-truth datasets and code to support future developments in improving IMS data analysis.

The general nuclear magnetic resonance analysis toolbox (GNAT) has established itself as a versatile free and open-source software suite for processing and analyzing NMR data. Here, we present a major...

An international team led by researchers at University of California San Diego and University of California, Riverside has developed a free, web-based platform designed to make public metabolomics data more accessible.