1
0
Entering edit mode
6.7 years ago
bruce.moran ▴ 880

Hi,

I am trying to run VEP, have installed it using the perl INSTALLER.pl which completed fine, using v79, for humans on build GRCh37. I have VCFs called with GATK. The error is for each line:

where 'P' is all chromosomes, 'L' all lines of VCF.

So I imagined the error was that Ensembl fasta was chromosomes named 'P', versus my own 'chrP', so I changed my VCF, still get error. Changed fasta, VEP creates a new index and still get warnings. So I thought I would ask. What format does VEP want? Google reveals nothing on this error. Sample of my VCF below:

Ensembl Variant Effect Predictor VEP Ensembl • 2.0k views
0
Entering edit mode

Hi Will,

files were there originally as you show, I did a clean reinstall and it is working now. Thanks for your help.

3
Entering edit mode
6.7 years ago
EnsemblWill ▴ 560

Probably there was an issue setting up the cache; if it reports they are all missing, then chances are the files genuinely aren't there.

If you used INSTALL.pl and the default parameters, you should see a bunch of files in each chromosome directory, e.g.

> ls \$HOME/.vep/homo_sapiens/79_GRCh37/Y/ | head
10000001-11000000.gz
10000001-11000000_reg.gz
10000001-11000000_var.gz
1000001-2000000.gz
1000001-2000000_reg.gz
1000001-2000000_var.gz
1-1000000.gz
1-1000000_reg.gz
1-1000000_var.gz
11000001-12000000.gz

Check that you have enough disk space where you are putting the cache files, then try running the installer again to get the cache.

Finally, VEP doesn't care whether the VCF has "chrY" or "Y", it changes this internally for you. However, the names in the FASTA file must be consistent with the cache file, so please try and use the cache and FASTA provided by Ensembl where possible.

0
Entering edit mode

Thank you Will, I had the same error, and restoring the cache (which was flawed somehow) definitely fixed it.