VEP missing refseq annotation in refseq cache?

Question

[michal.inbar]

Hi,

Emily_Ensembl

I'm using VEP102. cache homo_sapiens_refseq/102_GRCh37

For some reason I'm not getting the expected data when using --refseq cache.

When using --merged cache I do get the expected refseq data.

(I would use the merged cache, but then I cannot use --pick option and get a refseq result. see https://github.com/Ensembl/ensembl-vep/issues/790)

I'm also getting the expected data when using the online tool v103.

This is the command: (using --merged or --refseq)

vep --af --appris --biotype --buffer_size 500 --check_existing --distance 5000 --hgvs --mane  --protein --pubmed --merged --regulatory --species homo_sapiens --symbol --transcript_version --tsl --cache --dir_cache /usr/local/bioinfo/vep-data  --input_file testrefseq.vcf --output_file testmergedVEP.vcf --port 3337

The refseq output looks like this (for many different variants):

CSQ=C|intergenic_variant|MODIFIER|||||||||||||||||||SNV||||||||||||||||||||||||||||||||||||||||||||||

while the merged refseq returns different annotation for this position:

C|missense_variant|MODERATE|TNFRSF14|8764|Transcript|NM_003820.3|protein_coding|1/8||NM_003820.3:c.20G>C|NP_003811.2:p.Trp7Ser|320|20|7|W/S|tGg/tCg|rs200056967||1||EntrezGene||||||NP_003811.2|rseq_mrna_match||G|G||deleterious(0)|possibly_damaging(0.755)|||||||||||

This is the variant mentioned above, but it happens for all other variants I tested:

##fileformat=VCFv4.2                            
#CHROM  POS ID  REF ALT QUAL    FILTER  INFO
chr1    2488123 .   G   C   .   .   .

Thanks for your help :)

Answer 1

[Emily]

This works OK at our end. Possibly your cache hasn't downloaded properly. Could you check if you have the file /usr/local/bioinfo/vep-data/homo_sapiens_refseq/103_GRCh37/1/2000001-3000000.gz as this will give the output for your example. If that's there and accessible, could you please try a simplified run and tell us your result: perl vep -id 'chr1 2488123 . G C' --cache --dir $CACHE_DIR --port 3337 --refseq