Informations / Authors

Author	MOEHLIN Julien (Github, Gitlab)
Author	MOLLET Bastien
Author	COLOMBO Bruno Maria
Author	MENDOZA PARRA Marco (Github)
Team	SysFate
Email	mmendoza@genoscope.cns.fr

MULTILAYER TOOL

Changelog 2024

• The tool now detect automaticaly if the data are interstitial
• Fixed pattern detection in case of interstitial data
• Added a script to increase the number of pixels
• Improved several graphical elements

Description

Inspired by contextual pixel classification strategies applied to image analysis, we have developed MULTILAYER, allowing to stratify spatially-resolved transcriptome maps into functionally-relevant molecular substructures. For this, MULTILAYER applies agglomerative clustering within contiguous locally-defined transcriptomes (herein defined as gene expression elements or gexels), combined with community detection methods for graph partitioning.

Run

Launch :

python3 Multilayer.py

Dependencies

- Numpy

pip install numpy

- Matplotlib

pip install matplotlib

- Pandas

pip install pandas

- Scipy

pip install scipy

- Scikit-learn

pip install scikit-learn

- Seaborn

pip install seaborn

- Networkx

pip install networkx

- Louvain (from Community)

pip install python-louvain

- Python Imaging Library / PIL

pip install pillow

Only needed for UMAP

- Louvain (from Louvain)

pip install louvain

- Scanpy

pip install scanpy

if you use conda :

conda install -c conda-forge scanpy python-igraph leidenalg

MULTILAYER COMPRESSOR

Description

We developed module for compress data. Multilayer compressor is able to merge several gexels in one big gexel.

Run

Launch :

python3 Multilayer_Compressor.py -i input.tsv -o output.tsv -cx 100 -cy 100

Dependencies

- Numpy

pip install numpy

- Pandas

pip install pandas

MULTILAYER EXPANDER

Description

We developed module for extend data. Multilayer expander is able to divide one gexel into several gexels. n is the number of expansion.

Run

Launch :

python3 multilayer-expander.py -n 1

Dependencies

- Pandas

pip install pandas

VISIUM CONVERTER

Description

You need cellranger of 10xgenomics, as well as the datasets provided by the Visium platform; namely the matrix in h5 format (Feature / cell matrix HDF5), the feature information files (Feature / cell matrix), the spatial information data (Spatial imaging data), as well as our specialized script “visium converter.py” (dependencies: pandas package).

Run

First, use this command on “cellranger” to convert the h5 matrix to csv format: ./bin/cellranger mat2csv Feature/cell matrix_HDF5.h5 out_file_matrix.csv

Then, use our python script “visiumConverter.py” as following:

python3 visiumConverter.py -m out_file_matrix.csv -p spatial/tissue_positions_list.csv -g raw_feature_bc_matrix/features.tsv.gz -o matrix_multilayer.tsv –compressor

Dependencies

- Pandas

pip install pandas

ENRICHR CONVERTER

Description

A converter for Enrichr libraries. Once the library converted, you have to place it in the directory called 'GO DB'.

Run

Launch :

python3 enrichr_converter.py -i input.tsv

Dependencies

- Pandas

pip install pandas

DATA

All data available : (Article - data)

• Berglund & al - Prostate cancer data

• Asp & al - Heart development data

• Moncada & al - Pancreatic adenocarcinoma

• Rodriques & al - High resolution brain

• Ståhl & al - Breast cancer

• Ståhl & al - Mouse olfactory bulb

• Liu & al - Whole mouse embryo

TUTORIAL

Multilayer tutorial