SMIntegration Workflow: Spatial Metabolomics and Transcriptomics Analysis

# SMIntegration : Spatial Multi-omics Integration Platform ## Introduction **SMIntegration** is an innovative open-source platform for integrated analysis of spatial transcriptomics and metabolomics data. It integrates spatial pattern recognition, differential comparison, network construction, and functional annotation into a unified workflow. Designed to address key challenges in spatial multi-omics correlation analysis, SMIntegration enables researchers to explore gene-metabolite co-regulation mechanisms through an intuitive web interface, revealing spatial heterogeneity in tissue development and disease progression. ## Key Features ### 🔍 **Spatial Pattern Discovery** - Automated identification of spatial expression patterns for genes/metabolites - Moran's I correlation analysis between transcriptomic/metabolomic modules ### 🧩 **Pixel-level Clustering & Cell Annotation** - 4 clustering algorithms: Seurat-LV, Seurat-LM, Seurat-SLM, UMAP-kmeans - SingleR-based automatic cell type annotation with reference datasets - Custom cell type annotation support ### ⚖️ **Flexible Differential Analysis** - ROI selection via: - Interactive tissue imaging - Clustering results - Cell annotation mapping - Differentially expressed genes (DEGs) and metabolites (DAMs) detection - Group-specific gene-metabolite correlation networks ### 🧬 **Functional Analysis** - KEGG pathway co-annotation analysis - Fisher's exact test for DEG-DAM pathway associations ### 📊 **Interactive Visualization** - Gene-metabolite co-localization analysis - RGB overlay imaging for multi-feature visualization ## Getting Started ### 🌐 Online Access Access the live platform without installation: 🔗 [[[https://metax.genomics.cn/app/SMIntegration](https://metax.genomics.cn/app/smintegration)] ### 💻 Local Deployment (Docker) 🔗 [[[Docker Installation Guide](https://docs.docker.com/get-started/get-docker/)] ```bash # Pull the latest Docker image docker pull mzlabresearch/smintegration:v-1.0 # Run with ShinyProxy docker run -d -p 8787:3838 mzlabresearch/smintegration:v-1.0 ``` Please use a web browser to access: http://localhost:8787 ### 📥 Input Data Format SMIntegration requires two **feature matrices** as input files: **Key Requirements**: 1、Pre-aligned datasets: Spatial metabolomics and transcriptomics must share identical pixel coordinates and resolution. If registration is needed, you may use either the built-in alignment tool in the Upload tab or pre-align with [SpatialData](SpatialData.md). 2、Supported Formats: SMIntegration requires two **feature matrices** in TXT or RDS format: **TXT Format**: - Columns 1-4: Feature name (metabolite or gene), pixel coordinates (x, y), feature abundance (Intensity or MIDCount) - Each row represents one spatial pixel **RDS Format (Seurat object)**: - Spatial location information stored in meta.data (x, y) - Feature matrix stored in data@assays$Spatial (Format: Features as rows, spatial pixels as columns, values represent feature abundance) > 📖 Detailed formatting guide available in-app (Tutorial Panel → Data Preparation) ## Usage Workflow 1. **Upload Data**: Import feature matrix in **Overall Distribution Analysis** Panel 2. **Core Analysis**: - Pattern Recognition: Identify spatial expression modules - Pixel Clustering: Group pixels using 4 algorithms - Cell Annotation: Automated (SingleR) or manual mapping - Differential Analysis: Compare regions (manual/automatic ROI selection) - Network Construction: Build condition-specific gene-metabolite networks - Functional Analysis: Pathway mapping and co-annotation analysis - Visualization：Dynamic exploration of spatial distributions and co-localization patterns ## Example Data Test dataset: **Mouse brain coronal adjacent sections** 1. **Spatial Transcriptomics Data**: - Acquisition & Processing: Stereo-seq, 0.05μm resolution, aggregated to 50μm resolution - Content: 14,530 pixels × 500 genes 2. **Spatial Metabolomics Data**: - Acquisition & Processing: AFADESI (+) mode, 50μm resolution, spatially registered to identical pixel coordinates - Content: 14,530 pixels × 500 metabolites Access: Built-in dataset in SMIntegration (Tutorial Panel → Example Data) Raw data: NGDC OMIX Repository ID: [OMIX011674](https://ngdc.cncb.ac.cn/omix/release/OMIX011674) ## Community & Support Developed by Haoke Deng (denghaoke\@genomics.cn)\ Last update: 2026-02-04