Spatial transcriptomics maps gene expression while preserving the physical location of cells within tissue, revealing how genes are expressed across different regions. HPCBio can:

1. Transform sequence and image data to give spatial resolution of gene expression.
2. Normalize (e.g., SCTransform or log-normalization) expression counts, then reduce via PCA and UMAP/tSNE to enable visualization and clustering
3. Define spatial domains by clustering spots with similar expression profiles. Spatially-aware methods additionally factor in neighborhood context, identifying coherent tissue regions or “niches.”
4. Perform spatially variable gene detectionto identify genes whose expression varies significantly across space, which often correspond to tissue architecture, gradients, or functional zones
5. Optionally deconvolute cell types to infer the proportion of each cell type at every location using a paired scRNA-seq reference

Follow-up analyses include:

• Trajectory / pseudotime analysis — ordering spots along developmental or differentiation gradients (Monocle, PAGA)
• Cell-cell communication — inferring ligand-receptor signaling between neighboring cell types (CellChat, NicheNet, COMMOT)
• Multi-sample integration — aligning tissue sections across conditions or patients to enable differential spatial analysis
• Integration with H&E histology — overlaying pathologist annotations with expression data for tissue phenotyping
• Gene regulatory network inference — identifying transcription factor activity across spatial domains (pySCENIC)
• 3D reconstruction — stacking serial sections to model gene expression volumetrically
• Xenium/MERFISH single-cell resolution — moving from spot-level to true single-cell spatial resolution for finer analyses

Key Tools

• Seurat — R-based; widely used for preprocessing, clustering, and integration of spatial + single-cell data
• Squidpy — Python; spatial statistics, neighborhood graphs, and ligand-receptor analysis
• BANKSY — spatially-aware clustering using cell neighborhood averaging
• SpatialDE / SPARK-X — detecting spatially variable genes
• RCTD / SPOTlight / cell2location — cell type deconvolution from scRNA-seq references
• Giotto — end-to-end spatial analysis suite in R/Python
• BayesSpace — Bayesian clustering with spatial smoothing
• COMMOT / NicheNet — cell-cell communication inference using spatial proximity