Skip to content

Installation

The recommended way to install ipyrad2 is to use conda, which will not only install ipyrad2 but also all of its required external dependencies: fastp, samtools, bcftools, bedtools, bwa-mem2, vsearch, and mafft.

conda install ipyrad2 -c conda-forge -c bioconda

For developers: installing a development version

Set up an environment with all dependencies 

git clone https://github.com/eaton-lab/ipyrad2.git
cd ipyrad2
conda env create -f environment.yml -n ipyrad2
conda activate ipyrad2

Then install ipyrad2 itself:

pip install -e . --no-deps

Optional Analysis Dependencies

The core environment is enough for the main assembly workflow, but some downstream analysis commands require extra Python packages that are listed as optional dependencies in pyproject.toml.

The main optional analysis packages are:

  • scikit-learn
  • umap-learn
  • toyplot
  • toytree

You can install them either with conda:

conda install scikit-learn umap-learn toyplot toytree -c conda-forge

conda install ipyrad2[analysis] -c conda-forge

Verify the Install

A minimal check is:

ipyrad2 -h

If you are a returning ipyrad user, note that the source install also exposes a legacy-style entrypoint:

ipyrad2-classic -h

That command exists for the older parameter-file workflow, but the main documentation in this site focuses on the newer ipyrad2 subcommand interface.

Common Pitfalls

Forgetting to activate the environment

If ipyrad2 or one of its external binaries is “not found”, first check that the conda environment is active:

conda activate ipyrad2

Installing only the Python package

pip install . alone does not install the external tools used by trimming, mapping, and assembly. That is the main reason the docs recommend creating the conda environment first.

Missing analysis extras

Some analysis methods import optional packages lazily. That means the main install can succeed, but a method such as PCA or UMAP may later fail with a message about missing scikit-learn or umap-learn. Install those extras only if you need the corresponding methods.

Using the wrong shell or Python

If ipyrad2 starts but behaves like a different installation than expected, check which executable is being used:

which ipyrad2

That is especially useful if you have multiple conda environments or both editable and non-editable installs on the same machine.

Where to Go Next