Can your provide the Conda install?

[YongxinLiu]

When using this program, we still meeting many problems, such perl module, python module.

[VGalata]

I have to agree - it would be great to be able to install the tool via conda.

[nick-youngblut]

There's always hybridspades as an alternative: https://bioconda.github.io/recipes/spades/README.html#package-spades

[VGalata]

@nick-youngblut I know that there are alternatives which are easier to install. However, sometimes one has to use specific tools, e.g. for comparative studies and benchmarks.

[nick-youngblut]

Yeah, it's too bad that comparative studies almost never comment on the actual usability of software (eg., installation and dependency management in a complex data analysis environment), but at least there is some literature on that important aspect of bioinformatics (eg., https://doi.org/10.1186/s13059-019-1649-8)

[jsgounot]

Hi. Sorry to hear that you experienced difficulties to install OMS. There has been a transition of the project management and I didn't see your last comment. Packaging OMS for conda would be indeed an interesting feature, I will look to this. Until then, note that a docker file is available too.

[nick-youngblut]

It appears that all perl package dependencies are on bioconda, except for switch:

• bioconda::perl-file-which
• bioconda::perl-file-spec
• bioconda::perl-statistics-basic
• bioconda::perl-statistics-r
• bioconda::perl-getopt-long

...so a conda env yaml could be created to handle those dependencies.

@jsgounot it appears that you bundle a lot of the other dependencies in opera-ms:

All other required programs come either pre-compiled with OPERA-MS or are built during the installation process. Binaries are placed inside the tools_opera_ms folder:

You could unload the burden of having to maintain this software within opera-ms by relying on conda for the software installation. All of those dependencies should be available via conda

[jsgounot]

Indeed, however this will require several changes within OMS code, and therefore additional tests. But I agree that packaging OMS for conda would be a huge improvement, I will try to make this as soon as possible.

[jsgounot]

Until I'm able to provide a correct and clean conda package for opera-ms, here is a workaround for the python and perl dependancies with conda :

conda create -n operams python=3.9
conda activate operams
conda install perl-app-cpanminus
conda install -c compbiocore perl-switch perl==5.26.2
conda install -c bioconda perl-file-which perl-statistics-basic perl-statistics-r

or if you have mamba:

mamba create -n operams python=3.9
conda activate operams
mamba install perl-app-cpanminus
mamba install -c compbiocore perl-switch perl==5.26.2
mamba install -c bioconda perl-file-which perl-statistics-basic perl-statistics-r

I recommend to split the compbiocore and bioconda channels during installs, there are some overlaps and some packages on compbiocore didn't compiled on my machine. Once you've done that, you can clone opera-ms and build as indicated in the README

git clone https://github.com/CSB5/OPERA-MS.git
cd OPERA-MS
make
perl OPERA-MS.pl check-dependency

Don't forget to download the reference database.

I hope this might help some of you in the installation process.

Edit: Add emblab-westlake suggestion.
Edit 2: Specify python to version 3.9 as suggested in #76

[emblab-westlake]

I got an error when I use this command, and this command is available:
conda create -n operams
conda install -n operams -c compbiocore perl-switch perl==5.26.2
conda install -n operams -c bioconda perl-file-which perl-statistics-basic perl-statistics-r

[emblab-westlake]

conda install perl-app-cpanminus

[drelo]

Dear @emblab-westlake could you solve this? I followed those steps and then I installed all dependencies through mamba but got this error while executing make

This is the output
LOG2.txt

[jsgounot]

Dear @drelo, thanks you very much for your report, I create a new issue for this (#71). I identified the problem and should provide a fix today, sorry for this. I encourage you to create a new issue when your problem is not directly linked to other issues.

[drelo]

Thanks, my bad I thought it was an error of the recipe instead. I will follow the other thread. Thanks for your time.