Spatial Transcriptomics

We offer end-to-end spatial transcriptomics analysis using sequencing-based platforms such as 10x Genomics Visium and BGI Stereo-seq. Our pipeline integrates spatial coordinates with gene expression to reveal tissue architecture, spatial domains, and ligand-receptor interactions at cellular or subcellular resolution.

Workflow Summary

  ---
config:
    theme: 'base'
    themeVariables:
        fontFamily: 'verdana'
        fontSize: '25px'        
---
flowchart LR
    subgraph Preprocessing["`**Preprocessing**`"]
        direction TB
        Raw["Raw Reads"]
        Align["Alignment & Barcode Decoding"]
        Quant["Gene Expression Matrix"]
        Spatial["Spatial Coordinates Integration"]
        Filter["Quality Control & Filtering"]
        Raw --> Align --> Quant --> Spatial --> Filter
    end

    subgraph Basic["`**Basic Analysis**`"]
        direction TB
        Norm["Normalization"]
        HVG["Highly Variable Genes"]
        Reduce["Dimensionality Reduction"]
        Cluster["Clustering"]
        Vis1["Spatial Visualization"]
        Norm --> HVG --> Reduce --> Cluster --> Vis1
    end

    subgraph Advanced["`**Advanced Analysis**`"]
        direction TB
        Annot["Spatial Domain Annotation"]
        DE["Spatially Variable Genes"]
        Traj["Trajectory/Gradient Analysis"]
        LRI["Ligand-Receptor Interaction"]
        Deconv["Cell Type Deconvolution"]
        Vis2["Custom Visualization"]
        Annot --> DE --> Traj --> LRI --> Deconv --> Vis2
    end

    Preprocessing --> Basic --> Advanced

Preprocessing

  1. Raw Reads: Input FASTQ files from Visium or Stereo-seq platforms.
  2. Alignment & Barcode Decoding: Align reads and decode spot/cell barcodes using SpaceRanger (Visium) or SAW (Stereo-seq).
  3. Gene Expression Matrix: Generate spot-wise or bin-wise count matrices.
  4. Spatial Coordinates Integration: Integrate spatial metadata (tissue positions, resolution, spot/bin geometry).
  5. Quality Control & Filtering: Filter low-quality spots or bins based on read depth, gene count, mitochondrial content, etc.

Basic Analysis

  1. Normalization: Normalize expression values (e.g., SCTransform, log-normalization).
  2. Highly Variable Genes: Identify informative features across spatial units.
  3. Dimensionality Reduction: Apply PCA, UMAP, or t-SNE for visualization.
  4. Clustering: Group spots or bins into spatial domains or cell clusters.
  5. Spatial Visualization: Map gene expression and clusters back to spatial tissue images.

Advanced Analysis

  1. Spatial Domain Annotation: Annotate clusters based on marker genes or reference data.
  2. Spatially Variable Genes (SVGs): Identify genes with spatial expression patterns using methods like SpatialDE, SPARK, or FindSpatiallyVariableFeatures().
  3. Trajectory or Gradient Inference: Detect spatial gene expression gradients or tissue axes.
  4. Ligand–Receptor Analysis: Infer inter-cellular communication using CellChat, NATMI, or SpaTalk.
  5. Cell Type Deconvolution: Deconvolute spot-wise transcriptomes using scRNA-seq references (e.g., RCTD, Tangram, SPOTlight).
  6. Custom Visualization: Generate high-resolution overlays, spatial heatmaps, and pathway activity maps.

Example

Clustered spatial domains are overlaid on tissue images, highlighting gene expression patterns in anatomical context.

Spatially variable genes reveal localized transcriptional programs across tissue regions.

Deconvolution results show predicted cell-type composition at each spatial location using single-cell references.

Inferred ligand–receptor interactions illustrate spatial communication between regions or cell types.

For detailed options and customization, contact us.