Tutorial: Spatial gene expression analysis with SSAM

• Installation
  • A step-by-step guide
  • (Optional) Using sctransform for normalization
  • SSAM’s source code
• Data Preparation
  • Download VISp data
  • Load data into python
  • Transform Data
  • Create the SSAMDataset and SSAMAnalysis objects
• Creating the vector field
  • Prepare data for KDE
  • Run KDE
  • Local maxima search and normalization
  • Visualization
  • Normalization
• SSAM de novo analysis
  • Clustering of expression vectors
  • Cluster annotation and diagnostics
  • Visualisng the clusters
  • Cell type map
• SSAM guided analysis
  • Single cell RNA sequencing data
  • Cell-type gene expression signatures
  • Generate a guided cell-type map
• Clustering Local L-1 Maxima
  • Removing outliers
• Diagnostic plots
  • Example 1: a large cluster that can be easily annotated
  • Example 2: a large cluster that cannot be easily annotated
  • Example 3: a small cluster that is good
  • Example 4: a small cluster that is questionable
• Cluster annotation
• Visualisation of 2D gene expression embeddings (t-SNE and UMAP)
  • Running t-SNE
  • Running UMAP
  • Plotting embeddings
• Identifying tissue domains
  • Perform circular window sampling
  • Clustering domain signatures
  • Visualizing identified domains
  • Post-processing the identified domains
  • The final plot
• Cell-type composition analysis in tissue domains
  • Performing the cell-type composition analysis
  • Plotting the composition of each domain
  • Plotting the composition of the entire tissue
• Experimental features
• Cell-type classification using Adversarial Autoencoders
  • Mapping cell types using an adversarial autoencoder
• Segmenting the SSAM cell type map