Overview | Basic tile processing | tileSelection | tileGroup | Parallelization strategies | group_FUN | tileIterator | Manual streaming | With tileApply | setup_FUN for per-worker initialization | Iterator splitting | Choosing between them | Parallelization backends

1. Load the prepared SpatialGraph | 2. Tissue layout | 3. Find size-2 and size-3 motifs | 4. Where does a motif live in space? | 5. Enrichment under a global label-permutation null | 6. Region-stratified vs unstratified | 7. Top-enriched motif → associated genes | 8. Per-sample meta-analysis | 9. Per-niche motif distributions | 10. Niche-defining motifs in physical space | 11. Cross-niche bridging motifs | 12. Bar chart of the most common size-3 motifs | What's next

Overview | spatialTilePlan | pixelTilePlan | pointTilePlan | Coordinate modes: input and output | Same-mode: spatial → spatial | Same-mode: pixel → pixel | Cross-mode: spatial → pixel and pixel → spatial | freeTilePlan and quadtreePlan | Adaptive decomposition with quadtreePlan() | Choosing tile size and padding | Summary

Overview | The image and tile plan | Why tile groups? Edge tiles differ in size | Per-worker model initialization | Running inference on one group | Processing all groups and reassembling | Parallelization | Summary