[BioC] what is the threshold for R homology package

[Marc Carlson]

Hi Rachel,

That's a good question to be asking if you want to make use of these 
packages.  If you are looking at the inpararanoid packages, the bimap 
objects will only expose the mappings that inparanoid indicates are 
perfectly matched.  If that is too conservative for you, the database 
inside has all of the available data for mappings.  I am not really sure 
where you are getting 1804 from, but if you use the bimap you will only 
get the most conservative mappging.  So for example:

dim(toTable(hom.Hs.inpSCHPO))

Will only get you those mappings where the seed_status is 100%, which is 
why you get many fewer results from this.

But if you do this instead:
res = select(hom.Hs.inp.db, 
keys(hom.Hs.inp.db,keytype="SCHIZOSACCHAROMYCES_POMBE"), 
cols="HOMO_SAPIENS", keytype="SCHIZOSACCHAROMYCES_POMBE")
dim(res)

This interface will you get everything (regardless of what it's score 
was or wasn't).

If neither of these approaches is what you wanted, you can also query 
the DB under the hood which also has the seed_status and score for each 
match.


   Marc



On 11/21/2012 07:10 PM, Rachel [guest] wrote:
> I found no of homology pairs in R homology package is different from http://inparanoid.sbc.su.se/download/current/table_stats/
>
> For example, there are 1804 homo-sapien-Spombe pairs in Hom.Hs.inp.db, inparanoid website claims,
>
> 5615 sequences H.sapiens.fa have homologs in dataset S.pombe.fa
> 2986 sequences S.pombe.fa have homologs in dataset H.sapiens.fa
>
>
> May I know did you apply any cutoff for the homology pairs in R package?
>
>
>   -- output of sessionInfo():
>
> ..
>
> --
> Sent via the guest posting facility at bioconductor.org.
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor