Node reference

Spatial nodes.

Spatial transcriptomics end to end. Native loaders read a SpaceRanger folder (Visium, with H&E and scalefactors), one binned resolution of a Visium HD run, Xenium, MERSCOPE, CosMx and single-molecule transcript tables, or attach coordinates from a CSV. Then density QC, an interactive Plotly viewer with lasso / box / polygon selection, region grouping into obs labels, the spatial neighbour graph and squidpy statistics (neighborhood enrichment, co-occurrence, Moran / Geary / sepal autocorrelation, Ripley), Hotspot modules, BANKSY spatial domains, cellular neighborhoods, NNLS spot deconvolution, image features, and two ligand-receptor routes.

24 nodesCategory · spatialin 4 templates

Resolution matters

Xenium, MERSCOPE and CosMx are single-cell resolution; Visium and Visium HD are spot / bin based. Load Visium HD applies a bin-count guard before anything is read into memory — 8 µm bins are the practical starting point; a 2 µm run is tens of millions of bins and is not a bug when it is refused.

Every spatial statistic needs Spatial Neighbors first. Spatial Visualization then draws the overlays — 16 plot types, from an H&E overlay and transcript density to neighbourhood enrichment and ligand-receptor panels.

The Spatial Viewer keeps its zoom across re-renders, highlights a single group, and streams large tables in full (v0.25.0). Regions drawn with the lasso, box or polygon tools become an obs column through Spatial Grouping, which Subset Data can then filter on.

Common pitfalls

“No spatial coordinates.” Coordinates live in obsm['spatial']. Merge Cell Meta only writes obs columns and can never create them — use a native loader or Add Coordinates.

Ligand-Receptor (squidpy) needs the OmniPath intercell snapshot installed in the Resource Center, licence accepted. The shipped Spatial Full Analysis template stops at that node until you do; the other nodes still run.

sepal is grid-only. The sepal mode of Spatial Autocorrelation requires a regular grid (Visium / Visium HD) and refuses single-cell coordinates.

Load Transcripts

loadTranscriptssince v0.9.0-beta.41

Load single-molecule transcript locations (Xenium / MERFISH / seqFISH) into adata.uns["transcripts"]. Connect after Merge Cell Meta. Does not change adata.X; required for transcript dot overlays in Spatial Viewer.

Node LibrarySpatial

InAnnData

OutAnnData (uns["transcripts"] added)

Parameters (7)

Name	Type	Default	Description
`file_path`	file (.csv, .tsv, .txt, .parquet)	—	Transcript file. Parquet preferred for Xenium.
`gene_col`	text	`auto`	Gene column name.
`x_col`	text	`auto`	X coordinate column.
`y_col`	text	`auto`	Y coordinate column.
`qscore_col`	text	`auto`	Quality score column.
`min_qscore`	number	`0`	Min Phred Q-score (Xenium standard: 20).
`max_transcripts`	number	`0`	Cap (0 = no limit). 5000000 recommended for Xenium.

Add Coordinates

addSpatialCoordinates

Add spatial X/Y from CSV/TSV directly to obsm["spatial"]. Has its own loader — no Load CSV needed. Use this when you ONLY need X/Y. For full metadata, use Merge Cell Meta (which also handles obsm["spatial"]).

Node LibrarySpatial

InAnnData

OutAnnData (obsm["spatial"] set)

Parameters (7)

Name	Type	Default	Description
`coords_file`	file (.csv, .tsv, .txt)	—	Coordinates file.
`x_col`	obs column	—	X column (auto-detected from x/x_centroid/pxl_col_in_fullres).
`y_col`	obs column	—	Y column.
`id_col`	obs column	—	Cell ID column (empty = file index).
`delimiter`	select	`,`	Auto from extension.
`drop_unmatched`	checkbox	`false`	If True, drop cells whose id_col value is not present in adata.obs_names. If False (default), keep unmatched cells with NaN coordinates.
`library_id`	text	`spatial`	Stored in uns["spatial"].

Spatial Density QC

spatialDensityQCsince v0.9.0-beta.41

Filter low-density spatial regions (Xenium edges, bubbles, control probe noise) using a grid. Always filters — use Workspace modal for interactive threshold tuning. Defaults: grid_size=100µm, min_cells=2.

Node LibrarySpatial

InAnnData (with spatial coords)

OutAnnData (filtered)

Parameters (4)

Name	Type	Default	Description
`grid_size`	number	`100`	Grid cell size (µm).
`min_cells`	number	`2`	Strictly more than this many cells.
`x_col`	text	`x_centroid`	Fallback obs column for X.
`y_col`	text	`y_centroid`	Fallback obs column for Y.

Spatial Grouping

spatialGroupingsince v0.9.0-beta.39 (replaces spatialInteractiveViewer)

Lasso/box-select tissue regions and label them as obs categorical column. Replaces spatialInteractiveViewer. Polygon geometry persists with the workflow — survives preview subsampling. Pairs with Subset Data for region-specific analysis.

Node LibrarySpatial

InAnnData (with spatial coords)

OutAnnData (obs[label_col] categorical)

Parameters (4)

Name	Type	Default	Description
`label_col`	text	`spatial_group`	Output obs column.
`unassigned_label`	text	`Unassigned`	Label for cells outside regions.
`overlap_policy`	text	`last_wins`	"last_wins" or "first_wins".
`groups`	array	`[]`	Drawn region polygons (set via Workspace modal).

Spatial Viewer

spatialViewersince v0.9.0-beta.41

Read-only interactive Plotly WebGL scatter. Color by obs column or gene, select with lasso, box or polygon, highlight groups, inspect stats — all without modifying adata. 16 colormaps with live previews. Categorical/continuous mode toggle. To persist a selection, use Spatial Grouping.

Node LibrarySpatial

InAnnData (with spatial coords)

OutAnnData (pass-through)

Parameters (4)

Name	Type	Default	Description
`color_by`	text	—	Last-used obs column or gene (hidden state).
`mode`	text	`auto`	categorical/continuous/auto.
`cmap`	text	`viridis`	16 colormaps.
`marker_size`	number	`3`	Scatter marker radius (1-8).

Spatial Neighbors

spatialNeighbors

Spatial neighbor graph based on physical coordinates. Required for all spatial statistics.

Node LibrarySpatial

InAnnData (with spatial coords)

OutAnnData (obsp spatial_connectivities)

Parameters (8)

Name	Type	Default	Description
`coord_type`	select	`auto`	auto uses the spot grid for Visium (array_row/array_col present) and point coordinates otherwise; or force "generic" / "grid".
`n_neighs`	number	`6`	6 hex Visium, 10-20 single-cell.
`n_rings`	number	`1`	Rings for Visium grid.
`delaunay`	checkbox	`false`	Use Delaunay triangulation.
`radius`	number	—	Search radius (auto if null).
`spatial_key`	text	`spatial`	Key in adata.obsm for spatial coordinates.
`library_key`	obs column	—	obs column separating tissue sections. Empty auto-detects canonical obs['library_id'] or reuses a stored Spatial Neighbors key; if neither exists, the object is treated as one section.
`drop_na_coords`	checkbox	`true`	Drop cells with NaN/Inf coordinates before graph construction. False reports an error instead.

Nhood Enrichment

nhoodEnrichment

Test if cell types co-localize or avoid each other (squidpy nhood_enrichment).

Node LibrarySpatial

InAnnData (with spatial neighbors + clusters)

OutAnnData (uns["nhood_enrichment"])

Parameters (3)

Name	Type	Default	Description
`cluster_key`	obs column	`leiden`	Cell type column.
`n_perms`	number	`1000`	Permutations.
`seed`	number	`0`	Random seed.

Co-Occurrence

coOccurrence

Spatial co-occurrence at multiple distances.

Node LibrarySpatial

InAnnData (with spatial neighbors + clusters)

OutAnnData (uns["co_occurrence"])

Parameters (6)

Name	Type	Default	Description
`cluster_key`	obs column	`leiden`	Cell type column.
`interval`	number	`50`	Distance interval.
`max_distance`	number	`0`	Upper distance bound in coordinate units. 0 derives a common bound across sections; a positive value uses thresholds from 0 to this value.
`n_splits`	number	—	Parallel splits (auto if null).
`library_key`	obs column	—	obs column separating tissue sections. Empty auto-detects library_id or a stored Spatial Neighbors key; otherwise one coordinate system. Resolved sections are independent and cross-section pairs are excluded.
`max_pair_evaluations`	number	`500000000`	Advanced quadratic-work safety cap on the sum of n_section². Subset cells or regions before raising it.

Spatial Autocorrelation

spatialAutocorr

Moran's I, Geary's C, or sepal diffusion time — identifies spatially variable genes (SVGs).

Node LibrarySpatial

InAnnData (with spatial neighbors)

OutAnnData (uns["moranI"], uns["gearyC"], or uns["sepal_score"])

Parameters (7)

Name	Type	Default	Description
`mode`	select	`moran`	"moran" / "geary" work on any coordinates; "sepal" measures diffusion time and requires a regular grid graph.
`genes`	text	—	Empty = HVGs.
`n_perms`	number	—	null = analytical p-value (moran/geary only).
`max_neighs`	select	`6`	sepal only — lattice degree: 6 hexagonal (Visium), 4 square.
`sepal_n_iter`	number	`30000`	sepal only — max diffusion steps; genes that do not settle are reported as unconverged.
`sepal_max_genes`	number	`1000`	sepal only — cap on genes tested (0 = all). sepal costs ~0.1 s per gene.
`n_jobs`	number	`1`	Parallel jobs.

Load Visium (SpaceRanger)

loadVisium

Reads a whole 10x Visium SpaceRanger folder — counts, real coordinates, true scalefactors, and the H&E image.

Node LibrarySpatial

InSpaceRanger outs folder

OutAnnData (obsm['spatial'], obs['in_tissue'], uns['spatial'][library] with images + scalefactors)

Parameters (6)

Name	Type	Default	Description
`directory_path`	directory	—	SpaceRanger output folder (the one containing the count matrix and spatial/).
`counts`	select	`filtered`	'filtered' uses SpaceRanger's tissue-called barcodes; 'raw' loads the whole slide including background.
`filter_in_tissue`	checkbox	`true`	Keep only spots with in_tissue == 1.
`load_images`	checkbox	`true`	Read the tissue PNG. Coordinates load either way.
`library_id`	text	—	Blank = read from .h5 metadata, else the folder name.
`var_names`	select	`gene_symbols`	Gene symbols or Ensembl IDs.

Load Xenium

loadXenium

Reads a 10x Xenium cell-feature matrix with its aligned micron centroids, without pulling in whole-slide morphology images.

Node LibrarySpatial

InXenium output folder

OutAnnData (obsm['spatial'] micron centroids, obs cell metadata, uns['xenium'])

Parameters (4)

Name	Type	Default	Description
`directory_path`	directory	—	Xenium folder holding cell_feature_matrix.h5 (preferred) or the MEX directory, plus cells.parquet/csv.
`library_id`	text	—	Sample identifier. Blank uses the folder name.
`gex_only`	checkbox	`true`	Keep Gene Expression features only, dropping controls when feature types are present.
`var_names`	select	`gene_symbols`	Gene symbols or Ensembl IDs. Names are made unique.

Load Visium HD

loadVisiumHD

Loads one chosen square-bin resolution from a Visium HD run, with a bin-count guard applied before anything is read into memory.

Node LibrarySpatial

InSpace Ranger outs folder (or a square_XXXum folder)

OutAnnData (Visium contract + obs['bin_size_um'], uns['visium_hd'])

Parameters (9)

Name	Type	Default	Description
`directory_path`	directory	—	Space Ranger outs folder, or an exact square_XXXum bin folder.
`bin_size_um`	select	`8`	2, 8, or 16 µm. 8 is recommended; 2 µm needs an explicit opt-in.
`counts`	select	`filtered`	Canonical options: 'filtered' or 'raw'.
`var_names`	select	`gene_symbols`	Gene symbols or Ensembl IDs.
`library_id`	text	—	Name for the uns['spatial'] entry. Empty delegates to the native Visium loader's metadata/folder resolution.
`load_images`	checkbox	`true`	Read the tissue image. Required for H&E overlays and Spatial Image Features.
`filter_in_tissue`	checkbox	`true`	Keep only bins marked in_tissue == 1.
`max_bins`	number	`2000000`	Barcode ceiling checked before loading. Prefer a coarser bin size over raising it.
`allow_2um`	checkbox	`false`	Required opt-in for 2 µm bins. Does not bypass max_bins.

Load MERSCOPE

loadMERSCOPE

Loads Vizgen MERSCOPE cell-by-gene counts with their metadata, micron coordinates, and optional coordinate transformation.

Node LibrarySpatial

InMERSCOPE export folder + counts/metadata files

OutAnnData (obsm['spatial'] microns, obsm['blank_genes'], uns['merscope'])

Parameters (5)

Name	Type	Default	Description
`directory_path`	directory	—	Root MERSCOPE output folder.
`counts_file`	file (.csv, .csv.gz)	—	Cell-by-gene CSV/CSV.GZ filename.
`meta_file`	file (.csv, .csv.gz)	—	Cell metadata CSV/CSV.GZ with cell IDs and center_x/center_y.
`transformation_file`	file (.txt, .csv, .csv.gz)	—	Optional Vizgen coordinate transformation in the selected export's images folder. Accepts its existing images-relative name or the desktop picker's absolute path; Oreumics resolves and validates the target, rejects external/traversal paths, and passes Squidpy an images-relative path. Empty = no transformation.
`library_id`	text	`merscope`	Library identifier written to obs and provenance.

Load CosMx

loadCosMx

Loads flat NanoString CosMx exports, using slide-wide global coordinates as the canonical spatial system.

Node LibrarySpatial

InCosMx export folder + counts/metadata files

OutAnnData (obsm['spatial'] global pixels, obsm['spatial_local'], obs['fov'], uns['cosmx'])

Parameters (10)

Name	Type	Default	Description
`directory_path`	directory	—	Root CosMx output folder.
`counts_file`	file (.csv, .csv.gz)	—	Cell-by-gene CSV/CSV.GZ file.
`meta_file`	file (.csv, .csv.gz)	—	Cell metadata CSV/CSV.GZ file.
`library_id`	text	`cosmx`	Library identifier written to obs and provenance.
`fov_col`	text	—	FOV column override. Blank auto-detects fov / FOV / fov_id.
`cell_id_col`	text	—	Cell-ID column override applied to both files. Blank auto-detects.
`x_global_col`	text	—	Global X override. Blank auto-detects known global-pixel names.
`y_global_col`	text	—	Global Y override. Blank auto-detects known global-pixel names.
`x_local_col`	text	—	Local FOV X override. Local coordinates never replace the canonical global ones.
`y_local_col`	text	—	Local FOV Y override.

Spatial Graph Statistics

spatialGraphStats

Reports the observed contact matrix between clusters plus each cluster's network role on the spatial graph.

Node LibrarySpatial

InAnnData (spatial graph + cluster labels)

OutAnnData (uns['<cluster_key>_interactions'], uns['<cluster_key>_centrality_scores'])

Parameters (3)

Name	Type	Default	Description
`cluster_key`	obs column	`leiden`	obs column with cluster or cell-type labels (needs at least 2 observed groups).
`normalized`	checkbox	`true`	Row-normalize the interaction matrix. True reports each source cluster's edge fractions; false reports raw edge counts.
`weights`	checkbox	`false`	Off counts edges; on sums stored connectivity weights.

Spot Deconvolution (NNLS)

spotDeconvolution

Estimates the cell-type mixture inside each spot with a transparent non-negative least-squares fit against a labelled reference.

Node LibrarySpatial

InAnnData (spots), Reference .h5ad with cell-type labels

OutAnnData (obsm[key_added] proportions, obs['deconv_dominant_type'], obs['deconv_residual'], uns['spot_deconvolution'])

Parameters (9)

Name	Type	Default	Description
`reference_file`	file (.h5ad)	—	Labelled single-cell reference .h5ad.
`celltype_key`	text	`cell_type`	Reference obs column holding cell-type labels, with no missing values.
`spatial_source`	select	`X`	Expression source for spots. Canonical UI options: 'X' or 'layer'; backend also accepts 'raw'.
`spatial_layer`	layer	—	Layer name required when spatial_source='layer'.
`reference_source`	select	`X`	Expression source for reference cells. Canonical UI options: 'X' or 'layer'; backend also accepts 'raw'.
`reference_layer`	text	—	Reference layer name required when reference_source='layer'.
`min_cells_per_type`	number	`10`	Reference types with fewer cells are dropped; at least 2 must remain.
`min_shared_genes`	number	`100`	Minimum genes shared between reference and spots.
`key_added`	text	`deconvolution_proportions`	obsm key receiving the spot-by-cell-type proportions.

Spatial Image Features

spatialImageFeatures

Extracts per-spot intensity, histogram, and texture descriptors from the tissue image already attached to the data.

Node LibrarySpatial

InAnnData (obsm["spatial"] + attached hires/lowres image)

OutAnnData (obsm[key_added] feature table, uns['spatial_image_features'])

Parameters (10)

Name	Type	Default	Description
`library_id`	text	—	Image-bearing library. Blank auto-selects only when exactly one library has an image.
`image_key`	select	`hires`	Canonical options: 'hires' or 'lowres'. The key must exist in the selected library.
`features`	select	`summary`	summary, texture, histogram, or a comma-separated combination.
`spot_scale`	number	`1.0`	Multiplier controlling the image crop taken around each spot.
`histogram_bins`	number	`10`	Bins used when the histogram family is selected.
`texture_distances`	text	`1`	Comma-separated pixel offsets for texture features.
`texture_angles`	text	`0`	Comma-separated angles in degrees.
`key_added`	text	`image_features`	obsm key receiving the feature table.
`n_jobs`	number	`1`	Requested worker threads; must be at least 1 and is capped at 4 internally.
`max_image_megapixels`	number	`250`	Image-size safety cap. Use a lower-resolution image before raising it.

Spatial Hotspot

spatialHotspot

Finds spatially autocorrelated genes and groups them into coordinated local-expression modules using Hotspot.

Node LibrarySpatial

InAnnData (obsm["spatial"] + non-negative expression)

OutAnnData (var['hotspot_module'], obsm['hotspot_module_scores'], uns['hotspot_autocorrelations'], uns['spatial_hotspot'])

Parameters (13)

Name	Type	Default	Description
`layer_key`	layer	—	Non-negative expression source. Blank uses X.
`model`	select	`danb`	Canonical options: 'danb', 'bernoulli', 'normal', or 'none'. Use a count model only with integer-like counts.
`umi_counts_obs_key`	obs column	—	Optional obs column containing finite positive full per-observation UMI library sizes. Empty computes library size from row sums of the full selected expression matrix before gene filtering.
`weighted_graph`	checkbox	`false`	Use a distance-weighted spatial KNN graph.
`n_neighbors`	number	`30`	Spatial KNN size used to build the graph.
`neighborhood_factor`	number	`3`	Positive Hotspot neighborhood scale factor.
`approx_neighbors`	checkbox	`true`	Use Hotspot's approximate neighbor search.
`seed`	number	`0`	Random seed used to make Hotspot's approximate-neighbor construction reproducible.
`library_key`	obs column	—	obs section/sample column used as a safety boundary check. Empty auto-detects library_id or the stored Spatial Neighbors key. Multiple observed sections are refused; subset one section first.
`fdr_threshold`	number	`0.05`	FDR cutoff selecting genes for the pairwise correlation and module step.
`max_genes`	number	`500`	Cap on FDR-ranked genes entering the quadratic step (hard cap 1500).
`min_gene_count`	number	`10`	Minimum total selected-matrix expression per gene before autocorrelation testing.
`module_min_gene_threshold`	number	`20`	Minimum genes required to form a module.

Spatial Ligand-Receptor

spatialLigrecBivariate

Scores ligand-receptor pairs per spot against the spatial graph, mapping where signalling is co-located.

Node LibrarySpatial

InAnnData (with spatial neighbors)

OutAnnData (obsm['spatial_lr_local'], uns['spatial_ligrec'])

Parameters (8)

Name	Type	Default	Description
`resource_name`	select	`consensus`	Ligand-receptor database. 'consensus' (human) and 'mouseconsensus' (mouse) are the usual choices.
`local_name`	select	`morans`	Local statistic scoring co-localization of the ligand and receptor around each spot. Options: morans, cosine, pearson, spearman, jaccard, product, norm_product, masked_spearman.
`n_top_interactions`	number	`50`	Interactions keeping per-spot scores. The full ranking is always saved.
`nz_prop`	number	`0.05`	Minimum fraction of spots expressing a gene before its pairs are tested.
`n_perms`	number	`0`	Permutations for per-spot p-values. 0 = skip (much faster).
`mask_negatives`	checkbox	`false`	Keep only co-localization, zeroing avoidance scores.
`use_raw`	select	`auto`	'auto' uses adata.raw when it exists (it predates HVG selection, which routinely drops ligands and receptors), otherwise X. 'true'/'false' force the choice.
`seed`	number	`1337`	Random seed for permutations.

Spatial Domains (BANKSY)

spatialDomains

Segments tissue into contiguous domains by blending each cell's expression with its physical neighbours.

Node LibrarySpatial

InAnnData (with obsm["spatial"] + HVG)

OutAnnData (obs[key_added], obsm["X_banksy"], uns["spatial_domains"])

Parameters (10)

Name	Type	Default	Description
`spatial_weight`	number	`0.8`	BANKSY lambda. 0 = non-spatial baseline, 0.2 = spatial cell typing, 0.8 = tissue domains. Visium domains are the exception: 0.2 with k_geom 18.
`k_geom`	number	`15`	Spatial neighbours for the neighbour-mean block (18 for Visium).
`resolution`	number	`1.0`	Leiden resolution.
`key_added`	text	`spatial_domain`	obs column receiving the domain labels.
`use_agf`	checkbox	`false`	Add the azimuthal Gabor filter block (local expression gradient).
`use_highly_variable`	checkbox	`true`	Restrict to highly variable genes — strongly recommended.
`n_pcs`	number	`20`	PCA components taken from the BANKSY matrix.
`n_neighbors`	number	`15`	Neighbours for the clustering graph.
`library_key`	obs column	—	obs column separating tissue sections (empty = one section).
`random_state`	number	`0`	Random seed.

Cellular Neighborhood

cellularNeighborhood

Groups cells by the mix of cell types surrounding them, revealing recurring tissue microenvironments.

Node LibrarySpatial

InAnnData (with obsm["spatial"] + cell type labels)

OutAnnData (obs[label_col], uns["cellular_neighborhood"])

Parameters (7)

Name	Type	Default	Description
`celltype_key`	obs column	`leiden`	obs column holding cell-type or cluster labels.
`k_neighbors`	number	`10`	Nearest neighbours forming each window (published default 10).
`n_neighborhoods`	number	`10`	Number of neighborhoods for MiniBatchKMeans.
`label_col`	text	`cellular_neighborhood`	obs column receiving the CN labels.
`include_self`	checkbox	`true`	Whether the index cell counts in its own window (true matches Schurch et al.).
`library_key`	obs column	—	obs column separating tissue sections.
`random_state`	number	`0`	Random seed.

Ripley

ripley

Ripley's L/F/G statistics — measures spatial clustering patterns.

Node LibrarySpatial

InAnnData (with spatial neighbors + clusters)

OutAnnData (uns ripley results)

Parameters (9)

Name	Type	Default	Description
`cluster_key`	obs column	`leiden`	Cell type column.
`mode`	select	`L`	"L" / "F" / "G".
`n_neigh`	number	`2`	Neighbor rank used by F and G. Retained in the common contract for L but does not change L.
`n_simulations`	number	`100`	Confidence envelope simulations.
`n_observations`	number	`1000`	Number of random query points for the F function (G and F use every cell).
`max_distance`	number	`0`	Shared maximum radius in coordinate units. 0 uses the smallest sqrt(convex-hull area / 2) across sections.
`n_steps`	number	`50`	Number of equally spaced radii from 0 through max_distance; at least 3.
`seed`	number	`0`	Random seed. Stable section/group-specific generators make subsampling and simulations reproducible.
`library_key`	obs column	—	obs column separating tissue sections. Empty auto-detects library_id or a stored Spatial Neighbors key; otherwise one study window. Every resolved section receives its own convex hull and simulated null.

Ligand-Receptor

ligandReceptor

L-R interaction analysis using squidpy ligrec (CellPhoneDB-like permutations).

Node LibrarySpatial

InAnnData (with spatial neighbors + clusters)

OutAnnData (uns["ligrec"])

Parameters (4)

Name	Type	Default	Description
`cluster_key`	obs column	`leiden`	Cell type column.
`n_perms`	number	`1000`	Permutations.
`threshold`	number	`0.01`	P-value threshold.
`seed`	number	`0`	Random seed.

Spatial Visualization

spatialVisualization

Spatial scatter and analysis overlays (nhood, co-occurrence, ripley, autocorr, ligrec).

Node LibrarySpatial

InAnnData (with spatial coords)

OutPNG image

Plots (16)spatial_scatterhe_overlaysvg_panelrgb_coexpressiontranscript_densityneighbors_graphgraph_statscn_enrichmentdeconvolution_maphotspot_module_mapnhood_enrichmentco_occurrenceripleyspatial_autocorrligrecspatial_overview

Parameters (31)

Name	Type	Default	Description
`plot_type`	text	`spatial_scatter`	7 plot types.
`color_by`	text	—	obs column or gene.
`cluster_key`	text	`leiden`	Cluster column for overlays.
`section_view`	select	`auto`	For co_occurrence/ripley. Canonical options: 'auto', 'summary', 'pooled', 'section', or 'facets'. Auto selects summary for multiple sections and one section otherwise. Ripley maps pooled to summary because independent study windows cannot be pooled.
`section_id`	text	—	Exact stored section ID used when section_view='section'. Empty selects the first stored section.
`ripley_mode`	select	`auto`	For ripley. Canonical options: 'auto', 'F', 'G', or 'L'. Auto follows uns['ripley_latest']; an explicit mode resolves a stored result with that mode.
`show_null`	checkbox	`true`	For a single-section Ripley view, draw the per-group simulated null median. With show_interval=true, also draw its pointwise 2.5–97.5 percentile band.
`show_interval`	checkbox	`true`	Draw available interval bands: section q25–q75 for multi-section summaries, or the simulated pointwise band for a single-section Ripley curve.
`clusters`	text	—	Advanced comma-separated conditioning groups to draw. Empty auto-selects informative groups for co-occurrence and the first groups for Ripley; at most 6 co-occurrence or 12 Ripley groups are shown.
`deconvolution_key`	text	`deconvolution_proportions`	For deconvolution_map, exact dynamic obsm key written by Spot Deconvolution.
`deconvolution_cell_type`	text	—	For deconvolution_map, required exact cell-type proportion column. Missing/unknown values fail with the available columns.
`hotspot_module_column`	text	—	For hotspot_module_map, required exact column in obsm['hotspot_module_scores'], for example module_1. Spatial Hotspot reports module_score_columns.
`map_section_key`	text	—	For result maps, optional obs column separating tissue sections. Empty auto-detects library_id; multiple sections render as facets.
`map_section_id`	text	—	For result maps, optional exact section value. Empty renders all detected sections as facets up to the 12-section safety limit.
`vmin`	number	—	For result maps, optional lower color limit. Empty uses 0 for deconvolution and the observed minimum for Hotspot.
`vmax`	number	—	For result maps, optional upper color limit. Empty uses 1 for deconvolution and the observed maximum for Hotspot.
`cmap`	select	`viridis`	Colormap.
`size`	number	`1.0`	Point size.
`figsize`	array	`[10, 8]`	Figure size [width, height] in inches.
`alpha_img`	number	`1.0`	Tissue-image opacity for spatial_scatter when Visium-style image metadata is available.
`bw`	checkbox	`false`	Render an available tissue image in grayscale for spatial_scatter.
`spot_size`	text	—	Manual spatial point size. null uses max(2, size × 6).
`scale_bar`	number	—	Optional scale-bar length in the same coordinate units as obsm['spatial']; null disables it.
`flip_y`	checkbox	`false`	Apply an extra Y-axis flip in the direct spatial_scatter renderer.
`tissue_alpha`	text	—	Optional tissue-image opacity override; null uses alpha_img.
`source_groups`	text	—	For ligrec plots, optional comma-separated ligand-producing source groups. Empty = all.
`target_groups`	text	—	For ligrec plots, optional comma-separated receptor-bearing target groups. Empty = all.
`pvalue_threshold`	number	`0.05`	For ligrec plots, hide interactions above this p-value.
`font_size`	number	—	Font size. null = matplotlib default.
`title`	text	—	Custom plot title.
`publication_quality`	checkbox	`false`	Generate high-resolution publication-quality figure.

Visualization

Bulk RNA-seq