Skip to main content

Installation: standalone

Note

This is the recommended way to install Pangraph

Pangraph CLI is a self-contained single-file executable, which can be downloaded from the "Releases" page of the project and run directly. Here is a table with direct links to the latest versions for convenience:

x86_64arm64
Linuxgnu, muslgnu, musl
macOSdownloaddownload
Windowsdownload-
Note

There are 2 flavors of pangraph executables for Linux platforms - "gnu" and "musl".

The "gnu" flavor is generally faster but requires GNU libc >= 2.17 to be present on the system. Linux distros like Debian, Ubuntu, Red Hat, Fedora, Arch all come with GNU libc by default.

The "musl" flavor can be slower, but it makes no assumptions about libc. It is suitable for distros like Alpine, which have no GNU libc, or on older distros which have very old GNU libc.

All flavors require Linux kernel >= 3.2.

Note

On Unix platforms, don't forget to set executable file permission. You probably want to rename the executable and place it into $PATH to simplify its usage:

# Assuming ${HOME}/bin/pangraph is in $PATH
mv pangraph-x86_64-unknown-linux-gnu $HOME/bin/pangraph
chmod +x ${HOME}/bin/pangraph
pangraph --help

Installing from command line​

You can also download pangraph programmatically, for example to add it to your Docker image or a bioinformatics pipeline:

GNU, Latest Version​

curl -fsSL "https://github.com/neherlab/pangraph/releases/latest/download/pangraph-x86_64-unknown-linux-gnu" -o "pangraph" && chmod +x pangraph

GNU, Specific Version​

curl -fsSL "https://github.com/neherlab/pangraph/releases/download/1.0.0/pangraph-x86_64-unknown-linux-gnu" -o "pangraph" && chmod +x pangraph

MUSL, Latest Version​

curl -fsSL "https://github.com/neherlab/pangraph/releases/latest/download/pangraph-x86_64-unknown-linux-musl" -o "pangraph" && chmod +x pangraph

MUSL, Specific Version​

curl -fsSL "https://github.com/neherlab/pangraph/releases/download/1.0.0/pangraph-x86_64-unknown-linux-musl" -o "pangraph" && chmod +x pangraph

optional dependencies - mmseqs2​

When building graphs, in addition to the native minimap2, Pangraph can use mmseqs2 as an alignment kernel.1 mmseqs2 is able to find homology between sequences with up to around 30% sequence divergence, at the cost of higher computational time.

Use of this kernel requires installation of mmseqs2. This can be done for example using conda with:

conda install -c conda-forge -c bioconda mmseqs2
note

Pangraph was tested with mmseqs2 version 17.b804f.

Footnotes​

  1. This kernel can be selected with the -k mmseqs option when running the pangraph build command. ↩