Skip to content

himelmallick/Commingle

Repository files navigation

Commingle

This is a repository to create an attention based framework for diagonal integration of large single cell omics data

Requirements

The Python libraries required for executing the various phases of this pipeline are listed in the environment.yml file located in the corresponding folder. Please ensure that these dependencies are installed before running the pipeline.

Usage

Executing the Complete Pipeline

To execute the complete pipeline, run the script run_commingle.sh with the required arguments appended to the python main.py command for the corresponding phase. Modify line 4 of run_commingle.sh to abc or harmony to specify preprocessing using ABC or Harmony, respectively.

Executing Individual Phases

  • To execute one of the phase individually:

     $ cd <PhaseName>
     
     $ python main.py --<Argument1> <Value1> --<Argument2> <Value2> ...
    
  • For executing the Commingle code first time, execute the following command:

    $ python main.py --first_run
    
  • For executing the downstream code with attention, execute the following command:

    $ python main.py --attention
    

Note:

1.) The implementation currently supports .h5ad file as input.
2.) Please delete the existing saved_data and tmp folder in Phase_2-Commingle/ before training on new dataset.

About

This repository houses the Python codes for multiview diagonal integration using deep learning.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •