conda_env_instructions_for_mac.md

1. Install miniconda or anaconda with python version 3.8 or above.

2. Run the following commands for conda and pip

conda update --all
conda install r-essentials r-rgl
conda install pandas numpy scipy scikit-learn jupyter gsl tzlocal simplegeneric natsort h5py tqdm patsy llvmlite numba networkx joblib numexpr pytables seaborn statsmodels pip
conda install -c conda-forge python-igraph louvain leidenalg

pip install --upgrade pip
pip install MulticoreTSNE anndata anndata2ri fa2 gprofiler-official scanpy rpy2 scrublet

MulticoreTSNE is optional and just for speed, so leave it out if you run into problems installing it.

3. Start R and run the following commands:

install.packages(c('devtools', 'gam', 'RColorBrewer', 'BiocManager'))
update.packages(ask=F)
BiocManager::install(c("scran","MAST","monocle","ComplexHeatmap","slingshot", "DropletUtils"), version = "3.15")

The version of the BiocManager may depend on the installed R version (here R>=4.2).

Troubleshooting

MulticoreTSNE

A current workaround (as of 23/09/2022) to install MulticoreTSNE is to clone the git repository, adjust the setup.py and install manually as follows (solution provided by Lisa Barros):

git clone https://github.com/DmitryUlyanov/Multicore-TSNE.git

Then open the setup.py and replace self.cmake_args or "--", by self.cmake_args or "",.

Then install via

python setup.py install