Contribution/Software Version/Citation

Our automated workflow is built on top of Snakemake and follows a layered architecture of Python CLI + Snakemake workflow + scverse ecosystem. If you want to contribute code or add a new module, we recommend forking the repository and creating a dedicated feature branch.

GitHub Contribution and Module Extension

Project homepage and issue tracker:

Architecture Overview

  1. The command-line entry point (spatialsnake/command_line.py) parses arguments, loads configuration, and constructs the Snakemake command.

  2. The scheduling layer (spatialsnake/workflow/Snakefile) selects rule files according to option and runpipe.

  3. The execution layer (spatialsnake/workflow/rules/*.smk + spatialsnake/workflow/scripts/*) performs the actual analysis tasks.

  4. The environment and configuration layer (environment.yml + requirements.txt + requirements-extended.txt + spatialsnake/workflow/envs/*.yaml) defines the runtime environment and step-specific configuration.

This design decouples parameter management, workflow orchestration, and algorithm implementation, making it easier to extend the project while preserving reproducibility.

How Snakemake is Organized

In the current implementation, the command line automatically performs the following core steps:

  1. The Spatialsnake CLI is built on top of docopt. Users specify analysis options and run types via command-line arguments, which are parsed through the entry-point file command_line.py.

  2. After parsing, the CLI automatically selects the appropriate scheduling branch according to the specified analysis option and run type, and executes the corresponding Snakemake rules (Snakefile).

  3. The CLI automatically loads the default configuration file for the target stage, spatialsnake/workflow/envs/{option}.yaml. Users may also provide a custom configuration file via the --configfile argument.

  4. Input and output paths for each module are automatically resolved from the parameters defined in the configuration file. If you need to add a new analysis module, consult the path-resolution helper functions in the Snakefile and adapt them accordingly.

  5. If you wish to add a new scheduling branch, be sure to also update the parameter configuration logic for the module’s input and output paths. For a new platform branch, we recommend reading the data with a standalone Python script and bypassing the integration step initially, to avoid extensive logic refactoring.

Software Version

The table below summarizes the user-facing core software stack of Spatialsnake. Only major workflow and analysis modules relevant to spatial transcriptomics are shown here; low-level dependency packages are intentionally omitted. Versions were consolidated from the installation tutorial, dependency definition files, and the validated spatialsnake_env runtime environment.

Core runtime

Component

Version

Role in the workflow

Spatialsnake

--*

Command-line workflow framework for end-to-end spatial transcriptomics analysis

Python

3.12.11

Main runtime for the CLI, workflow logic, and Python analysis modules

R

4.4.0

Runtime for R-based downstream statistics, visualization, and communication analysis

Snakemake

9.8.1

Workflow scheduler and reproducibility engine
Core Python analysis modules

Package

Version

Main use in Spatialsnake

spatialdata

0.5.0

Unified object model for multi-platform spatial transcriptomics data

spatialdata-io

0.3.0

Import and export support for multiple spatial technologies

spatialdata-plot

0.2.11

Spatial visualization backend

scanpy

1.10.4

Core preprocessing, clustering, embedding, and marker analysis

anndata

0.12.0

Matrix and metadata container for single-cell and spatial analysis

squidpy

1.6.5

Spatial neighborhood, graph, and image-aware analysis

cell2location

0.1.5

Reference-assisted spatial cell type annotation

scvi-tools

1.4.0

Probabilistic deep-learning framework supporting annotation and latent modeling

pydeseq2

0.5.2

Differential expression analysis in Python-based comparison steps
Optional Python downstream modules

Package

Version

Main use in Spatialsnake

cellphonedb

5.0.1

Ligand-receptor communication analysis

cellcharter

0.3.5

Spatial clustering and niche organization analysis

liana

1.6.1

Consensus ligand-receptor inference framework

pyscenic

0.12.1

Gene regulatory network inference

torch

2.8.0

Deep-learning backend required by several probabilistic models
Core R analysis modules

Package

Version

Main use in Spatialsnake

Seurat

5.5.0

Spatial object handling, coordinate extraction, and downstream R analysis support

CellChat

2.2.0.9001

Spatially informed cell-cell communication analysis

clusterProfiler

4.14.0

Functional enrichment analysis

edgeR

4.4.0

Count-based differential analysis support

ComplexHeatmap

2.22.0

Publication-grade heatmap visualization

AnnotationDbi

1.68.0

Gene identifier mapping and annotation support

org.Hs.eg.db

3.20.0

Human gene annotation database for enrichment workflows

org.Mm.eg.db

3.20.0

Mouse gene annotation database for enrichment workflows

circlize

0.4.15

Circular visualization support for advanced downstream plots

schard

1.0.0

transform from zarr to Seurat

Citation

If you use Spatialsnake in your study, we recommend citing both the workflow framework and the project repository in the Methods section:

  1. paper:

  2. Spatialsnake project repository: https://github.com/l-zh007/spatialsnake