search

rmhubley / RepeatMasker

RepeatMasker is a program that screens DNA sequences for interspersed repeats and low complexity DNA sequences.
Other
214 stars 48 forks source link

Offline deployment install from Conda #233

Closed ptrebert closed 6 months ago

ptrebert commented 8 months ago

What do you want to know? What is the recommended way to deploy the Conda package of RepeatMasker (v4.1.5) on an offline cluster?

Helpful context Our cluster infrastructure is operated in "disconnected" mode, i.e. there is no general internet access. Software setups using Conda are possible via local mirrors of the most relevant Conda channels (bioconda, conda-forge etc.). After installing RepeatMasker v4.1.5 from bioconda, it looks like the default database file share/RepeatMasker/Libraries/Dfam.h5 is downloaded on-the-fly during the first start-up, which obviously fails on our infrastructure. Is there a command line option to point to this (type of) database file if stored in a central location on the cluster? The -lib parameters apparently expects a FASTA file.

Is there a particular genome assembly or organism your question is about? If possible, please provide a link to a publicly available assembly and/or a species name. No, general question.

Have you installed RepBase RepeatMasker Edition for RepeatMasker? No (?)

Thanks for your help.

Best, Peter

rmhubley commented 7 months ago

I definitely don't recommend Conda. We have had nothing but troubles with that system. I would recommend installing the program under a user account or have a system admin do it by hand. Another option is to use the TETools docker or singularity containers if your cluster supports them.

ptrebert commented 6 months ago

Thanks for the reply. Given that there is no general internet access on our cluster, the possibilities are limited and make manual tool setups usually quite painful. However, the hint about the TETools docker/singularity container was what I needed, that's the most straightforward solution. As a suggestion, maybe the existence of that tool container could be mentioned somewhere in the readme?