[Biojava-l] 3D viewing?

[Tom Oinn]

Egon Willighagen wrote:
> On Tuesday 11 January 2005 11:49 am, Tom Oinn wrote:
> 
>>Andreas Prlic wrote:
>>
>>>I am the person who contributed the biojava - structure classes. As it
>>>turns out I am a very happy Jmol user already for quite a while now!  :-)
>>
>>Seconded - Jmol is very cool and surprisingly easy to integrate if you
>>have PDB format data lying around somewhere in your code. I believe it
>>supports other structure formats as well although I've never used them.
> 
> 
> Hi Tom,
> 
> I'm actually one of the Jmol developers (though we should thank Miguel for the 
> excellent 3D rendering of proteins), but was actually refering to a tighter 
> integration of BioJava and Jmol...

Hi Egon,

Although I agree that in principle this is a reasonable thing to do I'm 
not convinced that it's worth it in this case - is there any information 
loss in converting to PDB format then reading back into JMol? If not 
then I'd leave it at that, there would be relatively few gains from 
being able to do so directly.

One thing that I believe systems like Spice take advantage of is the 
ability to include scripting instructions in the input to JMol; I 
haven't looked at the biojava 3d classes in detail but I would have 
thought they'd lack this functionality - it's nothing to do with 
representing the structure fundamentally. If you were to implement a 
direct Jmol view over the biojava classes you'd have to have some way of 
duplicating this functionality.

Having a relatively standard intermediate representation such as PDB 
format flatfiles is generally a good thing and makes it easier to link 
components in a loosely coupled fashion at the expense potentially of 
some efficiency. My thoughts are that the efficiency is no big deal in 
this case and that the convenience of the intermediate representation 
(not to mention that it already exists and works!) makes it the 
preferred option.

Cheers,

Tom