Install immuneML with a package manager

This manual shows how to install immuneML with pip in an Anaconda virtual environment.

Install immuneML

1. Create a directory for immuneML and navigate to the directory:

mkdir immuneML/
cd immuneML/

2. Create a virtual environment using conda. immuneML has been tested extensively with Python versions 3.7 and 3.8, but not 3.9. immuneML depends on PyTorch, which fails to install with Python 3.9. To create a conda virtual environment with Python version 3.8, use:

conda create --prefix immuneml_env/ python=3.8

3. Activate the created environment:

conda activate immuneml_env/

4. Install basic immuneML including Python dependencies using pip:

pip install immuneML

Alternatively, if you want to use the TCRdistClassifier ML method and corresponding TCRdistMotifDiscovery report, include the optional extra TCRdist:

pip install immuneML[TCRdist]

See also this question under ‘Troubleshooting’: I get an error when installing PyTorch (could not find a version that satisfies the requirement torch)

5. Optionally, if you want to use the DeepRC ML method and and corresponding DeepRCMotifDiscovery report, you also have to install DeepRC dependencies using the requirements_DeepRC.txt file. Important note: DeepRC uses PyTorch functionalities that depend on GPU. Therefore, DeepRC does not work on a CPU. To install the DeepRC dependencies, run:

pip install -r requirements_DeepRC.txt --no-dependencies

Testing immuneML

1. To validate the installation, run:

immune-ml -h

The output should look like this:

usage: immune-ml [-h] [--tool TOOL] specification_path result_path

immuneML command line tool

positional arguments:
  specification_path  Path to specification YAML file. Always used to define
                      the analysis.
  result_path         Output directory path.

optional arguments:
  -h, --help          show this help message and exit
  --tool TOOL         Name of the tool which calls immuneML. This name will be
                      used to invoke appropriate API call, which will then do
                      additional work in tool-dependent way before running
                      standard immuneML.

2. To quickly test out whether immuneML is able to run, try running the quickstart command:

immune-ml-quickstart ./quickstart_results/

This will generate a synthetic dataset and run a simple machine machine learning analysis on the generated data. The results folder will contain two sub-folders: one for the generated dataset (synthetic_dataset) and one for the results of the machine learning analysis (machine_learning_analysis). The files named specs.yaml are the input files for immuneML that describe how to generate the dataset and how to do the machine learning analysis. The index.html files can be used to navigate through all the results that were produced.