Mismatching annotations with M. musculus

[Darlene]

I ran GO-Elite a few times without running into any difficulty until I
started looking at the gene members of the terms reported by the
program. I updated the GO OBO database yesterday, and the program
parses a build with date 6/26/09. I used the Entrez Gene relational
database and all other defaults. However, I can't confirm some of the
gene annotations that the program reports. For example, one term
reported is "negative regulation of Ras protein signal transduction,"
with gene members Centd3, Csf1, Edg2. If I look any of these up in
Entrez Gene, the list of GO terms partway down the page does not
include "negative regulation of Ras protein signal transduction." If I
go to AmiGO and look up the term, those genes are not listed, and
conversely, if I look up the genes, the term is not listed. This term
is not the only one with this issue in my results, and it seems to
occur most often in the terms associated with smaller numbers of
genes. Any help in explaining this would be greatly appreciated.

[Nathan Salomonis]

Hi Darlene,
Amigo doesn't provide the gene-GO associations for GO-Elite, rather,
these associations are obtained directly from Ensembl and EntreGene.
As such, the best place to verify these annotations is (in the case of
EntrezGene) is the EntrezGene search page at NCBI.

I found your annotations there.
Best,
Nathan

Sent from my iPhone