Skip to content

Latest commit

 

History

History
53 lines (41 loc) · 2.18 KB

README.md

File metadata and controls

53 lines (41 loc) · 2.18 KB

singlecell-qtl

A workflowr project.

Setup

To ensure all contributors are using the same computational environment, we use conda to manage software dependencies (made possible by the bioconda and conda-forge projects). Please complete the following steps to replicate the computing environment. Note that this is only guaranteed to work on a Linux-64 based architecture, but in theory should be able to work on macOS as well. All commands shown below are intended to be run in a Bash shell from the root of the project directory.

  1. Install Git and register for an account on GitHub

  2. Download and install Miniconda (instructions)

  3. Clone this repository (or your personal fork) using git clone

  4. Create the conda environment "scqtl" using environment.yaml

    conda env create --file environment.yaml
    
  5. To use the conda environment, you must first activate it by running source activate scqtl. This will override your default settings for R, Python, and various other software packages. When you are done working on this project, you can either logout of the current session or deactivate the environment by running source deactivate.

  6. Initialize git-lfs and download latest version of large data files

    git lfs install
    git lfs pull
    

If there are updates to environment.yaml, you can update the "scqtl" environment by running conda env update --file environment.yaml.

Warning: If you are using RStudio, you need to ensure that it recognizes your conda environment. If you launch RStudio by clicking on an icon, it doesn't use the current environment you have configured in your shell. On a Linux-based system, the solution is to launch RStudio directly from the shell with rstudio. On macOS, running open -a rstudio . from the shell causes RStudio to recognize most of the environment variables, but strangely it does not set the correct library path to the conda R packages. Suggestions for how to fix this are welcome.