After installing the latest Spateo, following the tutorial cannot reproduce the results

[elegantmedstu]

Dear author,

I am trying to install the Spateo package and encountered an issue. I set up a conda environment with Python 3.9.0 and installed Spateo using pip install git+https://github.com/aristoteleo/spateo-release. The installation process did not report any errors. Here is the code I used for installation:

mamba create -n Spateo2 python=3.9 -y mamba activate Spateo2 pip install git+https://github.com/aristoteleo/spateo-release However, when following the tutorial (Spatially constrained clustering (SCC) with binning data), I encountered an issue. I downloaded the original h5ad file and ran the code line by line. When executing dyn.tl.louvain(adata_bin60, resolution=0.45), the resulting plot differs from the one in the official tutorial, lacking many clusters. I would like to know if this is expected behavior or if there is an issue with my setup. I would appreciate your guidance on this matter. Here is the code I used:

python

import spateo as st import dynamo as dyn import anndata

Read .h5ad file

adata_bin60 = anndata.read_h5ad('/home/lxx/project/stero空间组测试/cpl/02.bin/mousebrain_bin60.h5ad') adata_bin60 # Preprocessing st.pp.filter.filter_genes(adata_bin60, min_cells=3, inplace=True)

Normalization

dyn.pp.normalize_cell_expr_by_size_factors(adata_bin60, layers="X")

Linear reduction

st.tl.pca_spateo(adata_bin60, n_pca_components=30)

Identify neighbors (KNN)

dyn.tl.neighbors(adata_bin60, n_neighbors=30)

Louvain clustering

dyn.tl.louvain(adata_bin60, resolution=0.45)

st.pl.space(adata_bin60, color=['louvain'], show_legend="upper left", figsize=(4, 3), color_key_cmap="tab20") I would greatly appreciate your assistance in resolving this issue.

Thank you for your time and support.

Best regards, elegantmedstu

[github-actions[bot]]

This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 14 days

[Sichao25]

Hi, sorry for the confusion. The package is still under development, so functions and tutorials are updated frequently. Your clustering looks good to me.