Skip to content
Snippets Groups Projects

Clustering Analysis

Infer identity of myoblasts, at several development stages, through Seurat single-cell clustering analysis.

Description

Add in-depth description.


Installation

Prerequisites/Requirements

This project requires to install :

  • R version 4.2.3,
  • a "recent" version of Conda (> 23.1.0).

Conda is used to build the self-contained Python environment that hosts the following packages :

  • umap-learn version 0.5.3: used for building the umap embedding (instead of the default R package uwot),
  • leidenalg version 0.9.0 (or later): used for applying Leiden clustering algorithm when calling the Seurat::FindClusters function.

Note that the Conda environment is only required for running one of the preprocessing scripts: 9396.R, WP.R or 5hAPF.R.

Steps

  1. Download this repository or clone it if you wish to contribute.
  2. Build the R environment:
    • launch the R project ClusteringAnalysis.Rproj,
    • download the dependencies with the R command: renv::restore()
    • (optional) install the Conda environment: renv:::renv_python_conda_restore(".")

Usage

Once you have restore the R environment, you can run the script: test.R. If everything goes smoothly, the folder "results/pbmc" should be populated with several result files.

To perform the clustering analysis on the 9396, White Pupae or 5hAPF datasets :

  1. Create a folder named "data" at the project root and a subfolder of "results" where the figures and files will be stored,
  2. Copy the dataset in the "data" folder,
  3. Two options:
    • either run the corresponding preprocessing script to build the ".h5Seurat" object,
    • or, directly copy the corresponding ".h5Seurat" object in the "results" subfolder.
  4. Run the "analysis.R" with the following specifications:
    • the variable savedir should correctly point to the subfolder created at step 1,
    • the variable ident should point to a valid cluster resolution.

Contributing

Found a typo in the documentation? Interested in adding a feature or fixing a bug? Then by all means submit an issue or take a stab at submitting a pull request. If this is your first pull request, it may be helpful to read up on the GitHub Flow.

Pull Requests

When submitting a pull request:

  1. Clone the repo.
  2. Create a branch off of master and give it a meaningful name (e.g. my-awesome-new-feature) and describe the feature or fix.
  3. Open a pull request on GitHub.

License

This project is licensed under the CeCiLL License - see the LICENSE file for details.


Project status

Ongoing.