[ensembl-dev] apparent and low confidence orthology

[Julien Roux]

Thanks a lot Matthieu for the detailed answer.
Just one last question: is the "is_tree_compliant" score only 0 or 1, or 
can it take any value between 0 and 1?
Julien

On 3/13/15 5:38 PM, Matthieu Muffato wrote:
> Dear Julien
>
> API's is_tree_compliant() is the same as BioMart's confidence.
> It tells you whether the pair of genes should be considered as 
> orthologues when looking at the tree (i.e. is their LCA a speciation 
> node).
>
> apparent_ortholog_one2one is now ortholog_one2one with a "low 
> confidence" / is_tree_compliant=0. possible_ortholog have also been 
> merged into the standard orthology categories with the same "low 
> confidence" / is_tree_compliant=0.
> In your case, I think you should select all the "high confidence" / 
> is_tree_compliant=1.
>
> We don't have compute scores per homology, but only for the 
> duplication/dubious nodes. The formula only depends on the 
> distribution of species absence/presence under the two sub-trees, 
> regardless of the annotation of the other internal nodes (above or below)
>
> Hope this helps,
> Matthieu
>
> On 13/03/15 07:45, Julien Roux wrote:
>> Dear mailing list,
>>
>> I am wondering how to access information on the homology objects using
>> the Compara API. In particular I would like to be sure that an orthology
>> relationships I am looking at are not an "apparent" orthology
>> relationships (reciprocal gene losses after a duplication event)
>> A few years ago, this info used to be included in the description()
>> field (e.g., "apparent_orthology"), but it doesn't seem to be the case
>> any more.
>> In the documentation I have seen the function is_tree_compliant(). What
>> is it giving?
>> Also in Biomart, you can get an orthology confidence score [0 low, 1
>> high]. What does it mean and how is it calculated? How do you access it
>> using the API?
>>
>> Last question: is there a score difference between an apparent orthology
>> following a duplication node or following a dubious duplication node?
>>
>> Thnaks for you help
>> Julien
>>
>

-- 
Julien Roux
Marie-Curie postdoctoral fellow
Department of Ecology and Evolution, University of Lausanne, Switzerland
http://www.unil.ch/dee/home/menuinst/people/post-docs--associates/dr-julien-roux.html
Tel: +41 78 700 2931