This repository contains the code for all bioinformatic analyses shown in Neuwirth and Malzl et al. 2024.
Most code is presented in the notebooks/sctools package and should easily be executable after installing the accompanying conda environment
given in envs/scpython.yml.
Installation of the prvided environments is illustrated by the scpython environment. To install an environment you will first need to install a variant of Anaconda. We recommend miniconda for this.
After this you can simply type
conda env create -f envs/scpython.yml
This will prompt conda to install all necessary packages as they were used for the presented analyses. Please note that, to properly use the environment with JupyterLab
you will have to install the ipy-kernel which is done by typing
conda activate scpython
python -m ipykernel install --user --name scpython
The table below details which environments in the enclosed envs folder was used for which parts of the analyses
| envrionment | part of the analyses |
|---|---|
| scpython | general scRNA-seq analysis like integration and cell type annotation |
| scenic | gene regulatory network inference of skin diseases Tregs |
| recombat | batch effect correction prior to SCENIC GRN inference |
Note on scpython: To run Milo analyses with Python. milopy has to be installed manually from github source as described here. We used v0.1.0
Please note that the presented cell trajectory analysis was conducted in a separate R environment with R 4.3.0 and Bioconductor 3.17 in which Monocle3 1.3.7 was installed as described here
pySCENIC seems to be ill maintained (at least when it comes to the PyPI packages). At the time of this writing both pySCENIC and the arboreto packages used by pySCENIC did not run when installing them from their PyPI distribution. This was due to the version not being bumped after fixing the encountered bugs which prevents the update of the PyPI packages. Therefore, I recommend to install those packages from the clones git repositories for the code to run properly.
The bundled sctools is there for reproducibility reasons and corresponds to the snapshot of the code used throughout this study. Furthermore, it is only usable when located in the current working directory (which means notebooks have to reside in notebooks in order to work properly. In case you are looking for a more maintained and easier to work with library consider the sctools repository.