[Biojava-l] [Biojava-dev] How can I calculate the RMSD of Global Alignment?

[Andreas Prlic]

Hi Armita,

About calculating the global alignment with no gaps: This makes sense
if you work with highly similar proteins or e.g. if you run time
simulations for one protein and want to compare different points in
time.  I assume in your situation that every CA atom has a well
defined match in the protein to compare with. In this case going
through the full pairwise structure alignment procedure is not needed,
since you already know what the equivalent residues are. Equivalent
sets of Atoms can be compared using the SVDSuperimposer class.

Andreas




On Mon, Dec 29, 2008 at 1:44 PM, Armita Sheari <armita_sh at yahoo.com> wrote:
> Dear Andreas,
>
> Thanks for your answer. I saw the run of the example you had linked. As I found, all rmsd numbers which were calculated in the example were related to Local Alignments. Wheares I need to find the RMSD of Global Alignment (without any Gap). And we should consider all Alpha Carbons of the proteins to calculate the RMSD.
>
> Which parameters should I change in align method to reach the Global Alignment with NO Gap?
>
> Thanks again,
> ArmitaSh
>
>
> --- On Thu, 12/25/08, Andreas Prlic <andreas.prlic at gmail.com> wrote:
>
> From: Andreas Prlic <andreas.prlic at gmail.com>
> Subject: Re: [Biojava-dev] How can I calculate the RMSD of Global Alignment?
> To: armita_sh at yahoo.com
> Cc: "Biojava" <Biojava-l at lists.open-bio.org>, "biojava-dev" <biojava-dev at biojava.org>
> Date: Thursday, December 25, 2008, 7:20 PM
>
> Hi Armita,
>
> I agree the missing documentation for all the alignment parameters is
> a problem...   It is not exactly clear from your mail what is the
> problem you encountered. The code that you sent in principle looks
> fine. I suspect you want to select the "best" of the alternative
> alignments? In that case you can simply take the first, since the
> alignments come out sorted. An example can be run from here:
>
> http://www.biojava.org/download/performance/biojava-structure-example1.jnlp
>
> The alternative alignments are sorted according to their number of
> structurally equivalent residues. The rmsd in the different
> alternative solutions is always kept under a certain threshold (one of
> the parameters), since one of the strategies is to try to keep the
> rmsd constant, while maximizing the number of structurally equivalent
> residues. Similar results are clustered together in the same cluster.
> In the example shown above this allows to find the multiple matches
> between the four chains of hemoglobin and myoglobin.
>
> Andreas
>
>
>
>
>
> On Wed, Dec 24, 2008 at 6:02 PM, Armita Sheari <armita_sh at yahoo.com>
> wrote:
>>      Hi everyone,
>>
>>  I want to calculate the RMSD relevent to the Structural Global Alignment
> of two proteins. I have used the align method of the StructurePairAligner class
> with default parameters, and then I have calculated the RMSD using the getRmsd
> method of the AlternativeAlignment class. But it seems something is wrong.
>> I think I should change some parameters of align method which are defined
> in StrucAligParameters class. Unfortunately, I couldn't find any
> documentation that describes the parameters (not in the api nor in the source
> code).
>>
>> I would be thankful if you take a look at my code and let me know your
> opinion about which parameter(s) I should change.
>>
>> StructurePairAligner structurePairAligner = new StructurePairAligner();
>> structurePairAligner.align(structure1, structure2);
>> AlternativeAlignment[] alternativeAlignment =
> structurePairAligner.getAlignments();
>> ClusterAltAligs.cluster(alternativeAlignment);
>>
>> double minRmsd = 1000;
>> double rmsd = 0;
>> for(int i = 0; i < alternativeAlignment.length; i++)
>>            {
>>                rmsd = alternativeAlignment[i].getRmsd();
>>                if(rmsd < minRmsd) minRmsd = rmsd;
>>                rmsd = 0;
>>            }
>> return minRmsd;
>>
>> Thanks,
>> Armitash
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> biojava-dev mailing list
>> biojava-dev at lists.open-bio.org
>> http://lists.open-bio.org/mailman/listinfo/biojava-dev
>>
>
>
>
>
> _______________________________________________
> biojava-dev mailing list
> biojava-dev at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/biojava-dev
>