Hey 🙂 I noticed that you compiled a subset of the VEP cache for testing at some point but its based on the GRCh37 genome. Do you have any notes on how you downsampled the VEP cache? I want / need to do that for CI testing of #C05URH17X5Z and would love something similar to your setup

Hello! It’s been a while since I generated those files, so the details are a bit vague in my memory. If I recall correctly, I started with narrowing down the region to use (I picked all of chromosome 21 since it’s <50 Mb long) and used it as the pipeline’s default reference. Then, I selected the corresponding annotation from the VEP cache for that chromosome. If you navigate inside the VEP cache directory you will reach a directory with folders pointing to all the chromosome and contigs in the GRCh3(7|8), it should look like the attached screenshot. I kept only the annotation for chromosome 21 and ignored all other chromosomes and contigs. Are you only interested in the VEP cache data, or are you also planning to run plugins?

Thats already a very good hint! We just need the cache as we need it for another tool that we employ. I'll try and go ahead with your approach!

oh, that's smart

my tar.gz is still > 400MB because the all_vars.gz is much larger for the newer VEP release panic

ok, so using WBCel235 for testing was not too dumb of an idea actually

yes but that does not work for my cace since I need human data, I will now try and revert to an older cache and then we will see if that works - i guess an older cache is fine for this case

I was going to suggest using an older cache, and if I am not wrong we are using cache version 107 for tests in raredisease.

Sarek tests is human data

I guess a general problem is the GRCh38 instead of 37. It just has larger files ...