Re: [Treesoft-treefam] Clean vs Seed Trees ?

[Jean-Karim H. <jk...@sa...>]

Hi Sebastien,

Historically, the seed trees were derived from the PhIGs clusters and
used to create the families.
Now, seed trees should only appear for curated families and are used as
a constraint on building the clean trees for these families. The
families are now built using the previous version of Treefam as seeds.
Clean trees use only fully sequenced species while full trees use all
the other available sequences.
The difference also lies in the building process since DNA alignments
are available and used for clean trees but not for full trees.
So the choice depends on the requirements of your project.

In the case of  TF106228, there is clearly something gone wrong as the
genes of the seed tree appear in TF352211. I believe that TF352211
should have been mapped to TF106228 and a new family created for the 2
orphan worm genes but it seems that the reverse happened.
Jue, do you have any idea about this ?

Cheers

J-K

On Thu, 2008-09-25 at 10:45 +0200, Sebastien Moretti wrote:
> Hi
> 
> Following 'Instructions' link on family pages I thought that Clean trees 
> are clean because they use only sequences from sequenced species. And 
> they are not manually curated compared to seed trees.
> Most of the time clean trees are larger than seed trees (e.g. TF101001).
> 
> In some cases, e.g. TF106228, clean trees are smaller than seed trees. 
> Although only sequenced species are available in the seed tree AND in 
> the clean tree.
> 
> 
> So, what are the real differences between Clean, Seed and Full trees in 
> TreeFam ?
> 
> Only Clean trees in TreeFam B ?
> 
> What should be the best kind of tree for large scale phylogenetic studies ?
> 
> Thanks
>