Agentic and Foundation AI for Metabolomics and Multi-Omics in Biomedical Discovery

Metabolomics and other omics technologies are now capturing molecular phenotypes at unprecedented scale and resolution, generating spatial, single-cell, longitudinal, and multimodal measurements across diverse biological systems and disease contexts. However, transforming these complex datasets into reproducible biomedical knowledge remains challenging. Omics data are often high-dimensional, heterogeneous, sparse, and generated across different analytical platforms, experimental settings, and patient cohorts, complicating their integration, interpretation, and translation.Recent advances in foundation models, generative artificial intelligence (AI), and autonomous AI agents offer new opportunities to address these challenges. These approaches have the potential to integrate metabolomics and multi-omics data with molecular structures, biomedical literature, imaging, and clinical information, while supporting data interpretation, hypothesis generation, and experimental planning. By learning transferable representations across diverse datasets and coordinating complex analytical workflows, advanced AI systems may help transform descriptive molecular profiles into testable biological mechanisms, actionable biomarkers, and novel therapeutic hypotheses.Particular emphasis will be placed on studies that connect computational innovation with meaningful biomedical questions. Contributions addressing model generalizability across cohorts, populations, analytical technologies, and experimental settings are particularly encouraged, as are studies reporting uncertainty, interpretability, reproducibility, benchmarking, and appropriate experimental or clinical validation. The collection also aims to stimulate discussion on the opportunities and limitations associated with the adoption of increasingly autonomous and generalizable AI systems in metabolomics and multi-omics research.This Research Topic aims to bring together methodological and application-driven studies at the intersection of advanced AI, metabolomics, and molecular biosciences. It will explore how domain-specific and multimodal foundation models can learn biologically meaningful and transferable representations from metabolic and multi-omics data; how AI-agent and multi-agent systems can support and orchestrate quality control, annotation, data integration, analysis, literature-grounded reasoning, and experimental design; and how generative, graph-based, geometric, causal, and explainable AI approaches can contribute to robust biological discovery.This Research Topic welcomes Original Research, Methods, Reviews, Mini-Reviews, Perspectives, Technology and Code articles, and other article types that advance the development, application, evaluation, and translation of AI-enabled approaches for metabolomics and multi-omics research. Topics of interest include, but are not limited to:• Foundation models and self-supervised representation learning for metabolomics, spatial omics, and multi-omics• AI agents and multi-agent workflows for automated, reproducible data analysis, knowledge integration, and biological hypothesis generation• Multimodal integration of metabolomics with genomics, transcriptomics, proteomics, microbiome, imaging, molecular structure, and clinical data• Generative, graph-based, geometric, and causal AI for pathway inference, metabolic network modeling, and mechanism discovery• AI-driven biomarker discovery, patient stratification, disease prediction, and precision medicine• AI-enabled drug-target identification, virtual screening, drug repurposing, treatment-response prediction, toxicity prediction, and natural-product discovery• Explainable and trustworthy AI, including uncertainty quantification, benchmarking, data harmonization, privacy, fairness, reproducibility, and experimental or clinical validation Topic editors: Magdalene K Montgomery, Dr.Hongpeng Zhou Source: https://www.frontiersin.org/research-topics/85313/agentic-and-foundation-ai-for-metabolomics-and-multi-omics-in-biomedical-discovery