[BioC] Problems with homology packages

[Seth Falcon]

Hi Roberta,

On  4 Jul 2005, jmacdon at med.umich.edu wrote:
>> thanks for your help, it's working now.  I have a further question,
>> once I got the HomologeneIDs, I'd like to get the Accession numbers
>> for both human and mouse.  The HomologeneIDs that I find do not
>> match with the ones I can retrieve from the NCBI Homologene home
>> page, possibly because they are based on two different release. Do
>> you know where I could find HomoloGene and LocusLink database
>> release that where used in the package?
>
> For the HomoloGene release, see ?mmuhomology. I'm not sure if this
> was built using LL or Entrez Gene ids. Since it was built on 15 May,
> I think Entrez Gene ids would have been used, but the URL for
> HomoloGene in ?mmuhomology points to an old build that seems to use
> LL ids. Maybe Seth Falcon or Chenwei Lin can comment?

The homologene data provided by NCBI underwent a significant change
prior to the BioC 1.6 release.  As part of the change, the homology
data for a number of species that were previously included in the data
were removed.  The meaning of the homologene IDs also changed, I
believe.

For this release we decided it was best to stick with the homology
packages from the previous release --- they were rebuilt for
compatibility with R 2.1.x, but if I am remembering the events
correctly, the data in the homology packages is identical to what was
available in the 1.5 release.

Our team _is_ working on an updated set of homology packages that will
be made available in the development annotation repository soon.

Hope that helps explain what you are seeing.

Best,

+ seth