Recreate Figures from the scPortrait Manuscript

Notebooks

• 1. Figure 2
  • 1.1. Fig 2a H&E overview of tissue region
  • 1.2. Fig 2b close up of H&E, transcripts and fluorescent stains from a selected tissue area
  • 1.3. Fig 2c
  • 1.4. Fig 2d schematic overview of VitMAE training paradigm
  • 1.5. Fig 2e Generate tSNE Xenium All Cell Types
    • 1.5.1. tSNE of all cell types
    • 1.5.2. Annotate different tSNE regions with single-cell images
  • 1.6. Fig 2f Macrophages
  • 1.7. Fig 2g Visualize macrophage cells from different leiden clusters generated on the image space
  • 1.8. Fig 2h map macrophage leiden clusters into spatial context
  • 1.9. Fig 2i Differentially Expressed Genes between Macrophage Leiden Clusters
• 2. Figure 3
  • 2.1. Fig 3b SCimilarity Image Embeddings
  • 2.2. Fig 3c Prediction Error by Cell Type
    • 2.2.1. example images for each 10X celltype
  • 2.3. Fig 3d: Sankey plots to visualize cell type overlap
  • 2.4. Fig 3e
  • 2.5. Fig 3f
  • 2.6. Fig 3g
• 3. Fig_S2a Single-Cell Images
• 4. Fig S2c Visualize Representations
• 5. Figure S3
  • 5.1. Fig S3a: CODEX tissue overview
  • 5.2. Fig S3b: select CODEX channels in a small area of the tissue sample
  • 5.3. Fig S3c: single-cell images across all CODEX channels
  • 5.4. Fig S3d: OT/Flow Matching illustration
  • 5.5. Fig S3e: marker gene expression inferred by flow matching per OT assigned cell type
  • 5.6. Figure S3f: comparision of gene expression space measured using CITEseq vs inferred by flow matching
  • 5.7. Figure S3g: CD21 expression with Germinal Center (GC) regions highlighted
  • 5.8. Figure S3h: inferred TCL1A expression with Germinal Center (GC) regions highlighted
  • 5.9. Figure S3i: CD3 expression with T-cell rich regions highlighted
  • 5.10. Figure S3j: inferrred CD2 expression with T-cell rich regions highlighted
    • 5.10.1. Statistical testing for plots g,h,i.j
  • 5.11. Figure S3k: OT assigned cell types visualized in spatial context across tissue section
• 6. Figure S4
  • 6.1. Fig S4a Test Set Xenium Cell Types
  • 6.2. Fig S4b Transcriptome & Image Information Overlap
    • 6.2.1. Transcriptome Clusters on Images UMAP
    • 6.2.2. Transcriptome Clusters on Images TSNE
    • 6.2.3. Image Clusters on Transcriptome
  • 6.3. Fig S4c knn neighborhood overlap to quantify difference between transcriptome- and image-based embeddings
  • 6.4. Fig S4d Macrophages
  • 6.5. Fig S4e Map Macrophages onto Tissue
  • 6.6. Fig S4f tissue annotation
  • 6.7. Fig S4g Differentially Expressed Genes between Macrophage Leiden Clusters
• 7. Figure S5
  • 7.1. Fig S5a: quantify degree of overlap 10X cell type labels and SCimilarity transcriptome cell type labels
  • 7.2. Fig S5b: quantify degree of overlap SCimilarity transcriptome cell type labels and SCimilarity image cell type labels