[Feedback] Site Feedback: Variant effect predictor Web Tool is broken!

[feedback at gramene.org]

URL         : http://ensembl.gramene.org/Sorghum_bicolor/Variation/Explore?db=core;g=Sb03g038160;r=3:66091220-66091225;source=Morris 2013;t=Sb03g038160.1;v=S3_66091222;vdb=variation;vf=93377

Subject     : Site Feedback: Variant effect predictor Web Tool is broken!

Name        : Daniel Burkhardt

Email       : dburkhar at umass.edu

Organization: UMass Amherst / CSHL

Comments    : 

This is a missense SNP in Sorghum bicolor: 
3 66091222 66091222 C/A +

I can tell you its a SNP because Gramene tells me it's a SNP: http://ensembl.gramene.org/Sorghum_bicolor/Variation/Explore?db=core;g=Sb03g038160;r=3:66091220-66091225;source=Morris%202013;t=Sb03g038160.1;v=S3_66091222;vdb=variation;vf=93377

However, when I search for it using the VEP tool, I get this: 
Uploaded Variation	Location	Allele	Gene	Feature	Feature type	Consequence	Position in cDNA	Position in CDS	Position in protein	Amino acid change	Codon change	Co-located Variation	Extra
3_66091222_C/A	3:66091222	A	-	-	-	intergenic_variant	-	-	-	-	-	S3_66091222	-
Which is strange because when you view the SNP in the genome browser, it's clearly in the coding region of the protein.

In fact, EVERY single SNP I upload returns the "intergenic_variant" response. Here's a list of SNPs, NONE of which are intergenic, but ALL of which return intergenic via the web tool:

1	72047736	72047736	G/T	+
3	66088641	66088641	A/T	+
3	66088704	66088704	C/A	+
3	66089468	66089468	G/T	+
3	66091222	66091222	C/A	+
3	66092453	66092453	C/A	+
5	2718318	2718318	T/C	+
6	58815534	58815534	T/A	+
6	58816731	58816731	A/T	+
9	2560825		2560825	C/T	+
9	2560980		2560980	C/T	+
1	15926295	15926295	C/G	+
1	15938101	15938101	T/C	+
2	55729112	55729112	G/A	+
10	50233259	50233259	T/A	+
10	50233609	50233609	A/T	+

Interestingly, the API downloadable version of VEP works fine. I think this could be an issue with how VEP on the website connects with the Ensembl database.

http://www.warelab.org/bugs/view.php?id=4995