[ensembl-dev] Ensembl Variation DB table defs

[Stuart Meacham]

Dear list,

Just a quick note. I have download four tables from the latest variation 
DB (homo_sapiens_variation_60_37e) and noticed a quirk in the table 
definitions.

The variation table has an 'ancestral_allele' field defined as text 
(65536 characters) where the longest entry is a single character.

The variation_feature table has an allele_string field also defined as 
text of which the longest entry is 3838 characters (variation_feature_id 
= 28429746, an indel).

However the allele table (which I believe is supposed to have individual 
entries for the alternate alleles described in the variation_feature 
table) has an allele field defined as varchar(255) meaning any allele of 
more than that characters gets truncated. Variation feature with ID of 
28429746 is one such example, thus the allele described in that table is 
incorrect as it doesn't completely describe the indel in question. 
Obviously one can get the full allele by doing some simple text 
manipulation on the allele_string field in the variation_feature table, 
but the point remains.

Cheers

Stuart