Installation

Python

PyPI

fastdfe can be installed with pip:

pip install fastdfe

fastdfe is compatible with Python 3.11 through 3.13.

The SFS parsing is provided by the sfsutils package, whose input backends are optional extras that fastdfe exposes under the same names: vcf (the cyvcf2 dependency, for VCF files), zarr (the zarr dependency, for VCF-Zarr stores) and arg (the tskit dependency, for tree sequences). All of them are installed with:

pip install fastdfe[vcf,zarr,arg]

Conda

To avoid potential conflicts with other packages, it is recommended to install fastdfe in an isolated environment. The easiest way to do this is with conda or mamba:

mamba create -n fastdfe -c conda-forge fastdfe
mamba activate fastdfe

The optional input backends are not installed automatically with conda. zarr and tskit are available on conda-forge, while cyvcf2 is available on bioconda, so both channels are required:

mamba create -n fastdfe -c conda-forge -c bioconda fastdfe cyvcf2 zarr tskit

Alternatively, for reproducibility, the environment can be defined in a file environment.yml:

name: fastdfe
channels:
  - conda-forge
  - bioconda
dependencies:
  - fastdfe
  - cyvcf2
  - zarr
  - tskit

The environment is then created and activated with:

mamba env create -f environment.yml
mamba activate fastdfe

fastdfe is then imported with:

import fastdfe as fd

R

The fastdfe R package is installed from GitHub with:

devtools::install_github("Sendrowski/fastDFE")

Once the installation has completed, the package is loaded in an R session with:

library(fastdfe)

The fastdfe R package serves as a wrapper around the Python library, and draws its figures with ggplot2. Loading the R package declares the Python requirement, which reticulate resolves into a suitable environment the first time the module is loaded:

fd <- load_fastdfe()

fastdfe is compatible with Python 3.11 through 3.13.

The input backends are optional extras: vcf for VCF files, zarr for VCF-Zarr stores and arg for tree sequences. Only vcf is declared by default. Additional backends are declared by calling install_fastdfe() before the module is loaded:

install_fastdfe(extras = c("vcf", "zarr", "arg"))
fd <- load_fastdfe()

An existing Python installation can be used instead by installing fastdfe as described under the Python tab and selecting its environment before loading the module:

reticulate::use_condaenv("~/miniforge3/envs/fastdfe", required = TRUE)
fd <- load_fastdfe()

The R package documentation describes the available functions in more detail.