AI-Driven Multimodal Integration for Single-Cell and Spatial Multi-Omics

The rapid expansion of single-cell and spatially resolved technologies — spanning genomics, transcriptomics, proteomics, epigenomics, and metabolomics — has generated unprecedented volumes of heterogeneous biological data. While each modality offers a distinct lens into cellular identity and function, their true power lies in integration. Yet combining datasets across platforms, resolutions, and experimental conditions remains one of the field’s most pressing computational challenges.Artificial intelligence is emerging as the cornerstone of next-generation multi-omics integration. From graph neural networks and transformer-based architectures to generative models and contrastive learning frameworks, AI methods are enabling researchers to align modalities, resolve batch effects, impute missing measurements, and extract biologically meaningful representations at scale. At the same time, spatial omics adds a critical coordinate system — mapping molecular profiles to tissue architecture and revealing how cellular niches shape phenotype and disease.This Research Topic brings together cutting-edge work at the intersection of AI, single-cell biology, and spatial multi-omics. We welcome contributions across the full methodological and applied spectrum, including but not limited to:- Multimodal data integration frameworks for joint analysis of single-cell omics (e.g. scRNA-seq + scATAC-seq + CITE-seq)- Spatial transcriptomics and proteomics analysis methods and tools- Foundation models and large language models applied to single-cell and multi-omics data- Graph-based and deep generative models for cell-state inference and trajectory analysis- Cross-modality imputation and alignment under data sparsity and batch variation- Spatial domain identification and cell–cell communication inference- Benchmarking studies comparing AI integration strategies across platforms and tissues- Clinical and translational applications in cancer, immunology, neuroscience, and developmentThis topic targets computational biologists, bioinformaticians, and experimentalists developing or applying AI-powered tools to decode cellular heterogeneity and tissue organization. By consolidating methodological advances and real-world applications, this collection aims to define the state of the art and chart a roadmap for the field. Topic editors: Lantian Yao Source: https://www.frontiersin.org/research-topics/81651/ai-driven-multimodal-integration-for-single-cell-and-spatial-multi-omics