[Bioc-devel] Future plans for Homology annotations
Marc Carlson
mcarlson at fhcrc.org
Wed Oct 24 22:41:52 CEST 2007
Dear Bioc-developers,
I am considering a major upgrade to the homology packages and I would
like your input. The primary aim of these packages will be to provide a
means (in the format of an annotation package) for users to connect a
gene of interest from one species to it's homologs in another species.
This all sounds simple in principle, but of course there is an entire
field dedicated to exploring this question. So the more minds that have
input into this the better.
At present my plan is to produce some packages that represent the latest
inparanoid ortholog mappings (
http://inparanoid.sbc.su.se/cgi-bin/index.cgi ). Since inparanoid uses
only the longest translated transcript from each gene, their mappings
are effectively gene-centric. They also do their organism vs organism
comparisons in a pairwise fashion, which in practical terms means that
their comparisons have been run relatively recently and across a good
number of different organisms. These factors make inparanoid attractive
for finding orthologs between a pair of species quickly, and for
remaining relevant as the number of sequenced genomes gets very large.
Since inparanoid is not the only way to look at homology, these packages
would probably be called something like org.Hs.inp.db (human inparanoid)
or org.Mm.inp.db (mouse inparanoid) and would contain the comparison
data for the species in question vs all the other species where
inparanoid data was available.
The older homology packages used homologene (
http://www.ncbi.nlm.nih.gov/sites/entrez?db=homologene ) and so
inparanoid could represent a change. It is possible that some of you
might prefer homologene to inparanoid. If so, then perhaps we should
make revised packages to support that as well? Or perhaps you feel that
there is a compelling reason for not doing this? It is also still
possible that I might have missed something else so if you know of
something that you think should be included or made into a homology
related package of its own, please let me know what you think.
Marc
More information about the Bioc-devel
mailing list