SEU-3D

3D visualization and analysis plugin for spatial transcriptomics embryo data based on napari.

Features

3D Visualization: Render spatial transcriptomics data in 3D with napari
Tissue Filtering: Filter cells by tissue type, slice, germ layer, and XYZ range
Gene Expression Analysis: Single, dual, and triple gene expression coloring
Similarity Search: Find genes with similar expression patterns
Moran's I Spatial Autocorrelation: Compute spatial gene enrichment
Differential Expression: Identify tissue-specific marker genes
Annotation: Cluster and label cell populations in 3D space
Surface Reconstruction: Generate 3D surfaces for tissue regions (pyvista)
Automatic Color Mapping: Deterministic tissue colors from a discrete colormap

Installation

pip install seu-3d

Then open napari and look for "Load spatial transcriptomics data" in the plugins menu.

Environment

All runtime dependencies are declared in pyproject.toml and installed automatically via pip:

pip install seu-3d              # core functionality
pip install "seu-3d[surface]"   # + pyvista, for 3D surface reconstruction

Python >= 3.9

Data Format

The plugin reads .h5ad files. Two fields are chosen in the widget when loading; everything else is auto-detected by fixed names:

Field	Location	Required	Used for
(your choice)	`adata.obsm` key	Yes	Spatial coordinates. 3D `(x, y, z)` preferred; if 2D `(x, y)`, `obs['z']` is used as z
(your choice)	`adata.obs` column	Yes	Tissue / cell-type labels (coloring, filtering, diff-expression)
`z`	`adata.obs`	Only when the coordinate key is 2D	Z position of each cell
`slices`	`adata.obs`	Optional	Slice/sample filter (falls back to `orig.ident`)
`orig.ident`	`adata.obs`	Optional	Fallback slice/sample filter
`germ_layer`	`adata.obs`	Optional	Germ-layer filter
`x_flatten`, `y_flatten`	`adata.obs`	Optional	"Show Flatten" 2D projection

Missing optional fields simply disable the corresponding tab; the rest of the plugin keeps working.

Quick Start

import napari
viewer = napari.Viewer()
# Add spatial transcriptomics h5ad file via the napari plugin menu

Update Log

[2.0.0] — 2026-07-19

Packaging:
- Declare runtime dependencies in pyproject.toml (previously pip install seu-3d pulled in nothing)
- Restrict packages.find to seu_3d* so stray directories are never published
- Remove the unused, broken weiwei/ source tree and the dead _umap_selection module
- Single-source the version from seu_3d.__version__
Bug Fixes:
- Fix slice filtering never applying (obs['slice'] vs obs['slices'] mismatch)
- Fix differential-expression tissue selection reading only the first character of the tissue name
- Fix annotation coordinate matching failing on floats (cKDTree nearest-neighbour matching) and guard empty selections
- Fix saving annotations before annotating raised AttributeError
- Fix NaN colors for genes with constant expression
- Deterministic tissue colors (removed random shuffle)
- Re-filtering and XY preview now reuse their napari layers instead of piling up new ones
Performance:
- Remove redundant full AnnData.copy() in filtering and surface generation; filtering now applies one combined mask (a single copy per click instead of four)
- Annotations accumulate in one persistent copy (no per-click dataset copy, and repeated annotations no longer reset earlier ones)
- Move h5ad reading, Moran's I, surface reconstruction and similar-gene search to worker threads (no more GUI freezing)
- Sparse-aware cosine similarity and column means (no full-matrix densification)
UX:
- Gene-expression views reuse a single gene_expression layer instead of stacking a new view per click
- Dock panel width is bounded and scrollable; AnnData info box is read-only and height-capped
- Documented the expected obs/obsm field names (see Data Format above)
Code Quality: remove duplicate/unused imports, unify gene-expression coloring into shared helpers
Housekeeping:
- Centralize 2D->3D coordinate stacking in Embryo (written back to obsm)
- Add MIT LICENSE and unit tests for Embryo

[1.1.17] — 2026-05-30

First public release as a napari manifest plugin: h5ad loading, tissue/slice/germ-layer/XY filtering, 1-3 gene expression coloring, similar-gene search, Moran's I, differential expression, annotation, pyvista surface reconstruction

[1.1.6]

First release with napari manifest support

[1.0.4]

Fix error that plugin could not show

[1.0.3]

Fix napari.yaml

[1.0.2]

Forgot to upload napari.yaml

[1.0.1]

First release

Acknowledgements

License

Released under the MIT License, like the upstream sc3D project it is derived from.

Metadata

Release files for seu-3d 2.0.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for seu-3d 2.0.0
File	Size	Uploaded
seu_3d-2.0.0.tar.gz	18.7 kB	Jul 19, 2026	Details

Built distribution (wheel)

Table of built distributions (wheels) for seu-3d 2.0.0
File	Interpreter	ABI	Platform	Reset
seu_3d-2.0.0-py3-none-any.whl 16.5 kB Jul 19, 2026	Python 3	none	any	Details

Total release size: 35.2 kB

Release files / seu_3d-2.0.0.tar.gz

Download URL	seu_3d-2.0.0.tar.gz
Size	18.7 kB
Tags	Source
SHA-256 checksum How to use checksums	`9a3620418003cda92e9388d718c9ad8038d8e30254f1e050f82a2e5104bbc819`
BLAKE2b-256 checksum How to use checksums	`d5c90bf285f7199bbb6a7c23931ac6039bae43bc609d65cfa1a3d5bd66e1eedc`
Upload date	Jul 19, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.12.13`

Release files / seu_3d-2.0.0-py3-none-any.whl

Download URL	seu_3d-2.0.0-py3-none-any.whl
Size	16.5 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`2bbe7bbbd5b80372f63c176f672fdf46a8f4db8bd7d35511c33b35f8e831a190`
BLAKE2b-256 checksum How to use checksums	`1ea899e77d1ecf821babdc4b67331bd3ea4e4534350ef2db84dba97ba607fc85`
Upload date	Jul 19, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.12.13`