Spanve: A Statistical Method for Detecting Downstream-friendly Spatially Variable Genes in Large-scale Spatial Transcriptomics Data

Citation

@article{gpb.25.SpanveStatistical,
  title = {Spanve: {{A Statistical Method}} for {{Downstream-friendly Spatially Variable Genes}} in {{Large-scale Data}}},
  shorttitle = {Spanve},
  author = {Cai, Guoxin and Chen, Yichang and Chen, Shuqing and Gu, Xun and Zhou, Zhan},
  date = {2025-11-24},
  journaltitle = {Genomics, Proteomics \& Bioinformatics},
  shortjournal = {genom. proteom. bioinform.},
  pages = {qzaf111},
  issn = {1672-0229},
  doi = {10.1093/gpbjnl/qzaf111},
  url = {https://doi.org/10.1093/gpbjnl/qzaf111},
  urldate = {2025-11-25},
}

Analysis code for the paper is available at Evaluate directory.

Installation

• Install by pip (recommend):

pip install Spanve

• Install by source

if there are some problems with the pip installation, you can install it from source code.

git clone https://github.com/gx-Cai/Spanve.git # or download the zip file and unzip
cd Spanve
pip install -e .  # install in editable mode

• no install usage:

# install required packages
cd Evaluate/Softs
pip install spanve_requirements.txt
cp Spanve.py $your_path

Usage

cli usage

spanve --help

Usage: Spanve [OPTIONS]

Options:
  -i, --input_file PATH       input anndata file(h5ad file.)
  -r, --running_mode TEXT     running mode, default is f(c:cluster;
                              i:imputation; f:fitting)
  -s, --save_path PATH        save path
  -v, --verbose BOOLEAN       verbose
  -n, --njobs INTEGER
  -p, --preprocessed BOOLEAN  int preprocessed or not.
  --help                      Show this message and exit.****

command line usage can only run in a standard h5ad file, where there is a anndata.AnnData.obsm key named 'spatial'.

python usage

Quick Start

from Spanve import Spanve
adata = sc.read_h5ad('data.h5ad')
spanve = Spanve(adata)
# -- fitting for spatial genes
spanve.fit()
spanve.save('result.csv')
# -- spatial imputation
X = adata_preprocess(adata)
X_ = spanve.impute_from_graph(X[:,spanve.rejects])

Details

see tutorial notebook or html page.

Known Issues

• When numpy version is 1.24 or higher, you may enrounter the following error:

AttributeError: module 'numpy' has no attribute 'int0'

solution: downgrade numpy, or install by source code. [will fix in future release]