I am trying to install QIIME2 on a machine that already has most of the dependencies for QIIME2 installed. It is also a machine with somewhat limited disk space (I pay extra for more storage), and so consuming ~5GB storage to reinstall software that I have already installed is wasteful (and expensive).
Is there a recipe for installing QIIME2 without using miniconda?
Rob
Hi @linsalrob, thanks for reaching out! The large installation size is certainly unfortunate, but right now it is a bit of a necessary evil. Because QIIME 2 involves so many different dependencies, it is pretty important that we keep a close eye on the precise versions of those external tools, otherwise things might start causing trouble for users during analyses!
That said, it is technically possible to install just the QIIME 2 packages that you wish, on their own, but you will have to ensure that you have satisfied all of the dependencies yourself. You can use conda, and run an install command within your preexisting environment with all of your fulfilled dependencies there (you will want to instruct conda to use the right channel, too). Right now conda is a hard requirement for us, because it provides a consistent platform for our users to get third party binaries, such as mafft, without any extra work on their end. That said, if you didn’t want to use conda at all, you could install all of the dependencies listed here by hand, and then fetch the source code for each QIIME 2 plugin and run python setup.py install in all of them, but this sounds pretty painful to me.
Unfortunately, if you do attempt to perform an install of QIIME 2 that isn’t officially supported, it will be very difficult for us to provide support to you, because we are really only able to support the deployments that we have handcrafted. If you have any additional thoughts or ideas, we would love to hear them. Thanks! 
@linsalrob, I forgot to mention, if disk space is a problem, you might want to check out our pre-built Amazon Web Services AMIs — the Amazon platform is generally pretty competitive when it comes to pricing. Let us know if you have any questions. Thanks!
Another minor comment about disk space: you can try running conda clean --all -y to clear conda’s cache (e.g. old/unused packages and other build/install artifacts).
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.