Open cdiener opened 3 years ago
Hi @cdiener thanks for pointing this out! We hope to have a conda package available soon! Right now, I added some hints to the readme how to install all needed dependencies from conda without root rights.
Awesome thanks! A working docker or singularity image would also be an alternative if that is easier.
good idea thanks! we try to come up with some alternative installation in the near future :)
just as intermediate solution, we added a conda installation tutorial
Sure, no hurry. We packaged it as a docker image for our workflows for now: https://github.com/Gibbons-Lab/gapseq-docker. Though I noticed there are some dependencies that are not mentioned, like curl for instance.
@jotech any updates on creating the bioconda recipe? As illustrated with your conda installation tutorial, a conda recipe really helps developers (eg., of bioinformatics pipelines) install & maintain the dependencies of all software.
In regards to the conda install tutorial, a bit simpler approach:
conda env yaml file:
channels:
- conda-forge
- bioconda
dependencies:
- bash
- r
- perl
- parallel
- gawk
- sed
- grep
- bc
- git
- coreutils
- wget
- barrnap
- bedtools
- exonerate
- glpk
- hmmer
- blast
- bioconda::libsbml
- r-data.table
- r-stringr
- r-stringi
- r-getopt
- r-doParallel
- r-foreach
- r-r.utils
- r-sybil
- r-biocmanager
- bioconductor-biostrings
- r-jsonlite
- r-renv
Create conda env:
mamba env create -n gapseq -f gapseq_env.yaml
Additional R package install:
R -e 'renv::install(c("glpkAPI", "CHNOSZ"))'
In regards to the gapseq install instructions, I'm guessing that you meant:
bash ./src/update_sequences.sh
instead of:
src/./update_sequences.sh
Thanks @nick-youngblut. I agree that a conda env yaml file would clean up the instructions. I'll update the conda instructions in January.
Once this is available as a conda installation I'm going to add it as a module in my https://github.com/jolespin/veba pipeline. The publication just got accepted and should be out soon.
@jotech in order for a bioconda package to be effective you should make frequent releases (maybe, at least once a year). The current release (https://github.com/jotech/gapseq/releases/tag/v1.2) is more than two years old even though you had commits a few weeks ago (https://github.com/jotech/gapseq/commit/2d739a56a26a1eb9841b0b4999b3498452923406). Are there any plans to make a new release of gapseq that includes the changes to the development version?
The installation for gapseq requires root access to the machine which is often not the case for servers. It would be awesome if there was a conda package to facilitate a more accessible installation.