[Biojava-l] SVDSuperimposer

[Andreas Prlic]

Hi Raphael,

Thanks for posting the link to "LaJolla" (funny that you chose this
name, it is the part of the town where my university is located...).
This is very interesting and exactly the kind of applications we are
happy to see BioJava being used for.

Do you want to provide a link back to your project site at
http://biojava.org/wiki/BioJava:BioJavaInside and also add your paper
to the list of papers that are using BioJava?

About the  BioJava algorithm : If you have an example where the
BioJava structure comparison does not give the result you would
expect, could you send it to me so I can take a look ?

Thanks,
Andreas


On Fri, May 8, 2009 at 6:13 AM, Raphael André Bauer
<raphael.andre.bauer at gmail.com> wrote:
> On Thu, May 7, 2009 at 2:59 PM, Andreas Prlic <andreas at sdsc.edu> wrote:
>> Hi Michal,
>>
>> The SVD superimposer actually does the same Singular Value
>> Decomposition in both BioJava and BioPython. The difference is that
>> BioJava also contains a complete alignment algorithm that can
>> calculate a structure alignment, if the region of similarity between 2
>> proteins is not known a priori.
>>
>> See also here:
>> http://biojava.org/wiki/BioJava:CookBook:PDB:align  and
>> http://biojava.org/wiki/BioJava:CookBook:PDB:aboutalign
>
> Maybe it is interesting for you to know that we developed a generic
> algorithm for the structural alignment of RNA and proteins called
> "LaJolla". The reference implementation is built using BioJava and
> released as open source at http://lajolla.sourceforge.net. I think
> LaJolla is straight forward to use and modify if you know how to use
> Maven and the BioJava core classes.
>
> Btw. The algorithm for the alignment of proteins used by BioJava
> yielded sometimes strange results - but I did not do an exhaustive
> evaluation. But maybe LaJolla is an alternative in those cases.
>
> Additionally -- if the BioJava project is interested in integrating
> LaJolla into BioJava this is really not a problem -- but I do not know
> if it blurs the clean toolkit/library character of BioJava.
>
> We are open for all suggestions regarding the algorithm and also
> regarding a cooperation with BioJava...
>
> Thanks,
>
> Raphael
> @Michal: LaJolla is also partly made in Poland -- so mabye you some
> people by chance..
>
>>
>> Andreas
>>
>> On Thu, May 7, 2009 at 5:24 AM, Michal Lorenc <m.t.lorenc at wp.pl> wrote:
>>> Hello Andreas,
>>> SVDSuperimposer from BioPython works only with proteins which have the same
>>> length, but in BioJava it does not matter.
>>>
>>> Do you published how SVDSuperimposer in BioJava works or do you know where
>>> could I found more information?
>>>
>>> Thank you in advance.
>>>
>>> Best regards,
>>>
>>> Michal
>>>
>> _______________________________________________
>> Biojava-l mailing list  -  Biojava-l at lists.open-bio.org
>> http://lists.open-bio.org/mailman/listinfo/biojava-l
>>
>
> _______________________________________________
> Biojava-l mailing list  -  Biojava-l at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/biojava-l
>