Installation

SOPRANO is compatible with Linux and MacOS. It is primarily written in Python, but supplemented by some R, Perl, Bash, and bioinformatic utilities; such as bedtools.

Due to the variety of languages, SOPRANO is installed inside a conda environment. Note that this requires an existing Anaconda download, or alternative variant, such as Mamba; a highly efficient re-implementation.

---

Automatic installation

Users with Anaconda, or Mamba, available on their system can attempt an automated installation from their command line:

. setup.sh <flavour>

The flavour of the installation is optional, and may be omitted. Currently available options are:

• mpi - recommended for CLI users who user and HPC installations.
• dev - for developers; who are further recommended to pre-commit install hooks for standardized static analysis before CI.

Some users have reported issues with the automatic installation, in particular on macOS systems. If you experience problems, please refer to the manual installation. Pull requests in this area are highly welcome.

---

Manual installation

This installation guide assumes that users are in the repository root.

1. Create and activate a host conda environment
   

   conda create --name soprano
   conda activate soprano
   

   1. Mac
      

      conda config --env --set subdir osx-64
      conda env update ---file env.yml
      

   2. Linux
      

      conda env update --file env.yml
      

2. Install R GitHub package dependencies
   

   Rscript install_R_pkgs.R
   

3. Install Python package and corresponding dependencies
   

   pip install -e .[<flavour>]
   

4. Decompress transcript files
   

   gunzip -k data/aux_soprano/*.gz
   

Validate your installation

After completing either installation procedure, users are encouraged to run the installation tests, via

pytest -s tests/test_installation.py

All tests should pass.