A community effort to collect a curated set of analysis pipelines built using Nextflow.

nf-core

I am trying to use subworkflow <https://nf-co.re/subworkflows/vcf_annotate_ensemblvep/|vcf_annotate_ensemblvep> (and strictly not snpeff) in a workflow, and have the vep cache on s3 populated by annotation-cache.  It is hard to follow all the cache preparation in sarek, but at the end of the day it looks like the cache must be on local disk?
```def dir_cache = cache ? "\${PWD}/${cache}" : "/.vep"```
What is the simplest way from `vep_cache` directory on s3 to using vcf_annotate_ensemblvep?

So I think we managed to patch the setup in sarek

It was patched in <#C01NHS4CC91|> cf <https://github.com/nf-core/rnavar/blob/dev/subworkflows/local/annotation_cache_initialisation/main.nf>
to allow for other s3 like structure

cc <@UPC8CHDKQ> I'll be doing a PR to fix this right away

I see, but it appears way down in the module `dir_cache` is a local directory, even if `cache` is not

I see, you mean here: <https://github.com/nf-core/modules/blob/master/modules/nf-core/ensemblvep/vep/main.nf#L36>

the way I see it, all files input to Nextflow being symlinked in the process, it should be a local file anyway

With fusion and such system, it works quite seamlessly

And I just added an extra layer of IDs in the folder so that you don't try to copy the full cache, but only the one you need for the process (if you store multiple versions of cache / and/or species