[Bioperl-l] Bio::DB::Taxonomy root not present

Hilmar Lapp hlapp at drycafe.net
Fri Jul 8 03:05:53 UTC 2011


Right, but do you agree that looking for the label 'root' shouldn't be  
the only way to identify it?

	-hilmar

On Jul 7, 2011, at 10:17 PM, Brian Osborne wrote

> Hilmar,
>
> Instead of addressing the side issues address what I named as the  
> most salient issue: if we base this module's behaviour on NCBI's  
> taxonomy - and all data says this module should mirror NCBI's  
> taxonomy - then a node called "root" should exist since it exists in  
> NCBI's data files. Right?
>
> BIO
>
> On Jul 7, 2011, at 6:28 PM, Hilmar Lapp wrote:
>
>>
>> On Jul 7, 2011, at 3:22 PM, Brian Osborne wrote:
>>
>>> It's also true that having 5 trees instead of 1 is incorrect  
>>> scientifically
>>
>> That's a strong statement and I'm not sure I agree with this -  
>> let's keep in mind that these are taxonomies, not phylogenetic  
>> trees of all of life. Not every taxonomy has a node for "all of  
>> life" or for LUCA. For example, ITIS, one of the most widely used  
>> taxonomies if you're not dealing strictly with molecular data, does  
>> not - there is one "tree" for each kingdom of life. (Not that I  
>> want to recommend that as a good thing.)
>>
>> I agree with your programming awkwardness argument, though I would  
>> add  that looking for a specific label of a node to identify the  
>> root is always a bad (because fragile) thing to do. A better way to  
>> identify a root node would be parent undefined, or being the same  
>> as the node itself. If the code did that for each the 5 or so  
>> children of 'root', then the fake root could be removed.
>>
>> At the very least code in Bio::DB::Taxonomy or anywhere else should  
>> not assume that there is a single root for a taxonomy. Because  
>> there are taxonomies for which this really isn't the case.
>>
>> 	-hilmar
>>
>> -- 
>> ===========================================================
>> : Hilmar Lapp -:- Durham, NC -:- hlapp at drycafe dot net :
>> ===========================================================
>>
>>
>>
>>
>> _______________________________________________
>> Bioperl-l mailing list
>> Bioperl-l at lists.open-bio.org
>> http://lists.open-bio.org/mailman/listinfo/bioperl-l
>

-- 
===========================================================
: Hilmar Lapp -:- Durham, NC -:- hlapp at drycafe dot net :
===========================================================







More information about the Bioperl-l mailing list