Thanks a lot, Arek and Benoit! > %pos: defines the percentage of positivity (similarity) between both homologs. > We (compara) don't use this number in our pipeline anymore, but keep it > in the DB so users can use it. Does it mean that I'd be better relying on %id instead of %pos? (Will %pos be abandoned sooner or later?)
For the clarification of the task I'm trying to accomplish: I'm looking for the evolutionary conservation of regulatory elements in the promoters of orthologues genes. Species for comparison are chosen to be relatively close, like rattus norvegicus and mus musculus. And all of this is done for computationally predicted regulatory elements, not those experimentally detected. > If you want to have "good" orthologs and paralogs, we can advice you to > use the one2one and apparent_one2one. I did consult http://www.ensembl.org/info/data/compara/homology_method.html, and there it says that apparent_one2one "... can also be the results of a wrong gene tree topology and wrong duplication node labelling. ", so I was cautious with this homology type. Shouldn't I be cautious? What is the probability/occurrence rate of "wrong gene topology/wrong duplication node labelling"? > I am cc'ing Albert Vilella in charge of the genetrees in compara. more understanding of "good" orthologues genes would be definitely helpful! -- Sincerely yours, Bogdan Tokovenko, PhD student at the Laboratory of Protein Biosynthesis, Department of Genetic Information Translation Mechanisms, Institute of Molecular Biology and Genetics, Kyiv, Ukraine
