Installation

Documentation current for HybPiper version 2.3.0

---

HybPiper can be installed a number of different ways:

1) Installation using conda (strongly recommended)

If you don't already have conda installed, see here for instructions on downloading and installing Miniconda.

To avoid issues with incompatible conda package versions, we strongly recommend creating a new conda environment for HybPiper, e.g.:

conda create --name hybpiper

---

Important

IMPORTANT INFORMATION FOR MAC USERS WITH APPLE SILICON (M1/M2/M3 chips with ARM64 architecture).

There is currently no native ARM64 conda package for HybPiper that can be installed on recent Macs with M1/M2/M3 chips. Instead, users will need to create an x86 (i.e. the chip architecture used in older Intel-based Macs) conda environment and install HybPiper there. HybPiper will then run using Rosetta 2.

Note that you can do this even if you have the 'Apple M1' version of Miniconda installed. Use the following commands:

CONDA_SUBDIR=osx-64 conda create -n hybpiper
conda activate hybpiper
conda config --env --set subdir osx-64

Then, add the bioconda and conda-forge conda channels (see below), and install HybPiper with the command:

conda install hybpiper

---

Before installing HybPiper, make sure you have the conda channels bioconda and conda-forge added by following the instructions here. The commands required are:

conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge

Then, activate the environment and install Hybpiper:

conda activate hybpiper
conda install hybpiper

Alternatively, you can create the environment and install HybPiper with a single command, then activate the environment:

conda create -n hybpiper hybpiper
conda activate hybpiper

You should then be able to run the command hybpiper or hybpiper --help.

2) Manual installation of pipeline and dependencies

If you prefer, you can manually install the HybPiper dependencies, and then manually install HybPiper. Once the dependencies are installed and in your $PATH, clone the HybPiper repository from GitHub:

git clone https://github.com/mossmatters/HybPiper.git

Move in to the repo directory and install HybPiper using pip:

cd HybPiper
pip install .

You should then be able to run the command hybpiper or hybpiper --help.

NOTE:

Biopython

Biopython 1.80 is required as it contains bug fixes for the SearchIO module that HybPiper 2.0 uses to parse output from Exonerate.

Update 01/12/2022: Biopython 1.80 has now been released, and is available as a conda install here.

Prior to this release, HybPiper users needed to install an in-progress version by cloning the repository from https://github.com/biopython/biopython and following the instructions under the heading 'Installation from Source' here. Alternatively, the HybPiper conda package for all versions prior to 2.1.0 includes pre-release Biopython version 1.80.dev0; note that this causes problems if the user then installs Biopython 1.79 via conda in the HybPiper conda environment. The HybPiper conda package for version 2.1.0 now uses the conda Biopython 1.80 release package.

Python and SPAdes

The SPAdes assembler prior to version 3.15.4 is incompatible with Python 3.10 and 3.11, see here. The highest SPAdes version available as a macOS conda package is 3.15.2. This means that the macOS conda package for HybPiper is restricted to Python version 3.9 at the highest. This restriction will be removed once SPAdes >= 3.15.4 is released as a macOS conda package.

Update 29/05/2024: SPAdes version 3.15.5 is now available on Bioconda. Bioconda installs of HybPiper version >=2.1.7 will now use this SPAdes version or higher and a Python version higher than 3.9.

3) Installation using Docker

See the instructions on the HybPiper Bioconda page here

Dependencies

• Python 3.9 or later (see note), along with the Python libraries:
  • seaborn
  • matplotlib
  • pebble. The conda install can be found here
  • progressbar2. The conda install can be found here.
  • scipy. The conda install can be found here.
  • pandas
  • biopython 1.80 or later, see note.
  • psutil. The conda install can be found here.
• Exonerate 2.40 or later
• BLAST 2.9.0
• DIAMOND. The conda install can be found here.
• BWA
• BBtools. The conda install can be found here.
• SPAdes >=3.15.4
• GNU Parallel
• samtools 1.14
• MAFFT. New requirement for HybPiper version 2.1.0

Check installation

By default, HybPiper will check for the presence of all dependencies when the command hybpiper assemble is run. Alternatively, you can run the following command to check dependencies without running the assemble pipeline:

hybpiper check_dependencies