The last hint may help. Will have a deeper look. But I am pretty sure it's not a version problem. I got a local installation of GO mid 2011 and the results are similar to the newest Version (teste with various Symbols).
Hello everyone,
I am trying to get in touch with GO and/or Ensemble. First of all, I want to get alls GO Terms related to Gen. Here is an example for querieng PTEN in the GO DB:
SELECT
acc, gene_product.symbol
FROM
gene_product
INNER JOIN dbxref ON (gene_product.dbxref_id=dbxref.id)
INNER JOIN species ON (gene_product.species_id=species.id)
INNER JOIN association ON gene_product.id=association.gene_product_id)
INNER JOIN evidence ON association.id=evidence.association_id)
INNER JOIN term ON (association.term_id=term.id)
WHERE
gene_product.symbol = 'PTEN'
and species.genus='Homo'
and species.species='Sapiens'
and term.term_type = 'biological_process'
group by acc;
The Results are 56 Terms, perfectly equivalent to the amigo web interface. But when I query the Ensemble DB like this, for the PTEN protein:
select
*
from
translation
INNER JOIN object_xref ox ON (translation.translation_id = ox.ensembl_id AND ox.ensembl_object_type = "Translation")
INNER JOIN xref USING (xref_id)
INNER JOIN external_db ON (xref.external_db_id = external_db.external_db_id)
INNER JOIN ensembl_ontology_66.term ot on (dbprimary_acc=ot.accession)
where
db_display_name = 'GO' and stable_id = "ENSP00000361021"
and ot.ontology_id=1
group by dbprimary_acc;
I get 42 GO Terms for biological process.
May one of you has a hint for me.
Thanks,
Mario
//EDIT:
I got a bit deeper into it and did 5 things:
- Use Amigo, searching PTEN
- Use SQL Query searching PTEN
- Use BioMart: Search for ENSP00000361021
- Use Ensemble v66 with query for ENSP00000361021
- Use Ensemble v66 with query for PTEN
1 and 2 are identical 3,4 and 5 are identical
the 42 GO terms are a subset of the 56. Is amigo may linking to other species?
2 answers
I believe this is a simple problem of version mismatch. The latest Ensembl release is from February 2012 while you are looking at GO from April 2012. Ensembl GO annotations are not updated in between releases, so in Ensembl you are looking at the annotations as they were sometime in January or so.
Note that Ensembl projects GO annotations across species using orthology predictions. There are specific rules for this and the new annotations are tagged with the appropriate evidence code (IEA).
I hope this helps
Nope, it's for mas data analysis. Need to solve it in SQL
Log in to answer this question.