This is a tutorial to:

1. download and install conda – the environment and package management tool
2. build a conda environment for computational biophysicist

Download and installation

It is recommended to use either miniconda or miniforge (mamba). The mamba has become more popular due to its quick speed at resolving dependencies.

access the two packages from the following links and choose one you like:

1. miniconda
2. miniforge/mamba

Follow their instructions to install them. Personally I use mamba. All commands you run use conda can be run by simply replacing conda with mamba.

Manage environment

After installation, the mamba or conda comes with an initial base environment. Of note, we almost never mess with the clean base environment. Instead, we create our own with:

mamba create -n bio python=3.12

# OR 

conda create -n bio python=3.12

For the python version, you usually needs to decide on what version is compatibale with most packages you are going to install in the created environment. If you are not sure, try not using the latest and the oldest python versions. For example, the current latest python is 3.13 as the time of writing this note. I would go for 3.10, 3.11 or 3.12.

We can check and remote an environment with:

# check all envs you have created
mamba env list

# remove environment
mamba deactivate # if you are currently inside this environment
mamba remove -n [env_name] --all

Configure a environment for computational biophysicists

mamba activate bio  # always activate the environment before installing any packages
mamba install -c conda-forge numpy pandas matplotlib scipy
pip install jupyterlab
pip install seaborn biopython mdtraj mdanalysis

# Optional: openmm
mamba install -c conda-forge openmm
# if you have nvidia GPU cards
mamba install -c conda-forge openmm cuda-version=12.0
# of note, the cuda-version should match with your system nvidim-smi cuda version
# you can check with the command: nvidim-smi

# Optional: psfgen
mamba install -c conda-forge psfgen

The biopython, mdtraj and mdanalysis are packages quiet often used for PDB and MD trajectory analysis. seaborn and matplotlib are popular packages for scientific plot.

You can further configure the environment according to your own needs.