The easiest way how to setup a computational environment for RASE is using Bioconda.
- Python 3
- ProPhyle
- ETE 3
- PySAM
- GNU Make or equivalent
- GNU parallel
- Ghost Script
- Pandas
- SnakeMake
- SAMtools
- R
- R OptParse
- GCC 4.8+ or equivalent
- Seqtk
- zlib
We recommend to create a separate software environment (here called rase
):
conda create -n raseenv \
prophyle ete3 pysam snakemake-minimal samtools parallel r-optparse pandas seqtk
The environment can then be activated by
source activate raseenv
To avoid failures on very long reads, we recommend upgrading ProPhyle to a corrected version:
source activate raseenv
pip install --upgrade git+git://github.com/prophyle/prophyle.git@b55e026
All the dependencies can also be installed without BioConda.
Many of these packages are distributed using standard package systems such as APT.
apt-get install build-essential python3 zlib1g-dev r-base r-cran-optparse ghostscript
All the Python packages (ProPhyle, PySAM, ETE 3, and Snakemake) can be installed using PIP:
pip3 install prophyle pysam ete3 snakemake
At some systems, the R package distributed by Bioconda might not be properly built and would display messages such as
dyld: Library not loaded: @rpath/libintl.9.dylib
Referenced from: /Users/user/miniconda/envs/rase/lib/R/lib/libR.dylib
Reason: image not found
Abort trap: 6
The solution is then to create the raseenv
environment without r-optparse
, and
to install R and the OptParse package manually.
ETE 3 library, which is used for tree plotting, internally depends on QT and requires using an X-Server. This becomes problematic especially on virtual machines. For instance, on Ubuntu-based machines this can be solved by installing several additional packages:
apt-get install xvfb libqt4-dev libgl1-mesa-dev libglu1-mesa-dev xauth xfonts-base
and then prepending the following string to commands for building the database.
xvfb-run --server-args="-screen 0 1024x768x24 -noreset" \