VEP v88 - Do not output any PolyPhen prediction terms with "--humdiv"

[henrikstranneheim]

I have used v88.8 with the 88 cache and the PolyPhen prediction term with “—humdiv” seem to still be missing from the cache, see below. I thought this was resolved in the 88 cache or am I doing something wrong?

See:

perl variant_effect_predictor.pl -v -i example_GRCh37.vcf --offline --assembly GRCh37 --dir_cache /mnt/hds/proj/cust003/develop/modules/miniconda2/envs/mip4.1/ensembl-tools-release-88.8/cache --force_overwrite --polyphen b --humdiv
#----------------------------------#
# ENSEMBL VARIANT EFFECT PREDICTOR #
#----------------------------------#

version 88

By Will McLaren (wm2@ebi.ac.uk)

Configuration options:

assembly           GRCh37
core_type          core
dir                /mnt/hds/proj/cust003/develop/modules/miniconda2/envs/mip4.1/ensembl-tools-release-88.8/cache
dir_cache          /mnt/hds/proj/cust003/develop/modules/miniconda2/envs/mip4.1/ensembl-tools-release-88.8/cache
dir_plugins        /home/henrik.stranneheim/.vep/Plugins
force_overwrite    1
host               ensembldb.ensembl.org
humdiv             1
input_file         example_GRCh37.vcf
offline            1
polyphen           b
port               3306
species            homo_sapiens
stats              HASH(0x58aa638)
verbose            1

--------------------

2017-05-16 13:28:13 - Read existing cache info
2017-05-16 13:28:13 - Auto-detected FASTA file in cache directory
2017-05-16 13:28:13 - Starting...
2017-05-16 13:28:13 - Detected format of input file as vcf
2017-05-16 13:28:13 - Read 173 variants into buffer
2017-05-16 13:28:13 - Reading transcript data from cache and/or database
[===============================================]  [ 100% ]
2017-05-16 13:28:14 - Retrieved 3097 transcripts (0 mem, 3162 cached, 0 DB, 65 duplicates)
2017-05-16 13:28:14 - Analyzing chromosome 21
2017-05-16 13:28:14 - Analyzing variants
[===============================================]  [ 100% ]
2017-05-16 13:28:14 - Calculating consequences
[===============================================]  [ 100% ]
2017-05-16 13:28:14 - Analyzing chromosome 22
2017-05-16 13:28:14 - Analyzing variants
[===============================================]  [ 100% ]
2017-05-16 13:28:14 - Calculating consequences
[===============================================]  [ 100% ]
2017-05-16 13:28:15 - Processed 173 total variants (86 vars/sec, 86 vars/sec total)
2017-05-16 13:28:15 - Wrote stats summary to variant_effect_output.txt_summary.html
2017-05-16 13:28:15 - Finished!
(mip4.1) [henrik.stranneheim@rastapopoulos variant_effect_predictor]$ grep -c PolyPhen variant_effect_output.txt
1

[willmclaren]

The intended fix for v88 GRCh37 cache slipped through the cracks, so the data are missing for this version. v89 is definitely fixed; the cache has already been produced and tested ready for release soon.

In the meantime, you should be able to use an older version of the cache without much issue. v86 has the humdiv data present, and is largely the same in terms of its data content. You may have VEP install an older version with e.g.

$ perl INSTALL.pl -v 86 -s homo_sapiens -y GRCh37 -a c
$ ./vep -cache_version 86 [etc]