homebrew Vs bioconda

Question

[hamid.gaikani]

Hi,

I'm wondering if homebrew and Bioconductor are doing both the same job or not? I've seen two different introductory videos. One tutor uses Bioconductor to download packages (and also creating new environments), and the other one used homebrew. Now, I want to download another package, and I don't know if I should "brew install X" or "conda..."

Could someone tell me what the difference is? And does it really matter which one to use?

I am very novice to bioinformatics. So sorry if this is a dumb question.

Answer 1

[Alex Reynolds]

The bioinformatics part of Homebrew (brewsci/homebrew-bio) had some non-trivial issues until recently. The bioconda channel seems to get more love and attention. Packaging software for easy distribution is not easy.

My gut feeling is that whatever gets your job done correctly and in a reproducible way is probably doing decent science.

Use whatever tools in the toolbox to that end. Keep a manifest of packages and versions that you use, and how they were compiled. Packagers are no substitute for learning the command line.

[hamid.gaikani]

Got it! Thanks very much for the reply. It is really helpful!

Answer 2

[Juke34]

If you know both you can use any of them. But if you don't know any of them and you have to choose one, I would go for conda, the community is bigger as well as the number of packaged tools.

In your case I would bay attention specificaly to how to install add-on R packages.

With conda you install the r packages from conda, and they provide an easy way to create r packages from CRAN when the recipe is missing:
https://github.com/bgruening/conda_r_skeleton_helper
https://conda.io/projects/conda-build/en/latest/user-guide/tutorials/build-r-pkgs.html

I don't know how things have evolved for Homebrew but in the past you had to install R packages from sources.