SMODEL

Requirements

Before running SMODEL, please ensure the following MATLAB toolboxes are installed:

• Statistics and Machine Learning Toolbox
• Deep Learning Toolbox

Usage

Clone this repo

git clone https://github.com/liying-1028/SMODEL.git
cd SMODEL

Code description

• Data_processing.py: Data preprocessing
• Run_SMODEL_Simulated_data.m: SMODEL for simulated spatial triple-omics datasets
• Run_SMODEL_real_data.m: SMODEL for real datasets
• visualization.ipynb: Clustering visualization

Example

Take the dataset "simulated spatial triple-omics" as an example

• Step 1: Prepare the data. The SMODEL takes the preprocessed data and the base clustering results (poolsget) as input. For data preprocessing, please refer to the following script:

  Data_processing.py
  

After preprocessing, the data should be organized and saved into a MATLAB .mat file, which will be directly used by SMODEL. We provide an example MATLAB script, DataMat.m, for generating this file.

• Step 2: Main running pipeline:

  Run_SMODEL_Simulated_data.m
  

• Step 3: Visualization:

  visualization.ipynb
  

Note:

• To reduce the time required for reproduction, we have uploaded our preprocessed data into the folder data. Users may directly use these files and skip Step 1 if desired.
• To reproduce the result, you should use the default parameters. For new datasets, we recommend starting with the default configuration (see Run_SMODEL_real_data.m). Additionally, you may adjust the parameters according to the unique characteristics of your dataset to attain the best possible outcomes.