VMD-L Mailing List

From: Tomek Wlodarski (tomek.wlodarski_at_gmail.com)
Date: Tue Nov 24 2009 - 05:17:59 CST

Dear Boris,

Thank you for reply!
I know that sequence alignment based on structure superposition can have
more than one solution like you said.
However when I get one I can manually adjust obtained sequence alignment,
especially in those sites where there is one residue in the middle between
two residues of another structure to obtained better conservation of
residues.
Frankly I do not know how precisely Multiseq is creating sequence alignment
based on superposition of structures done by Stamp algorithm...
I just wanted to use the same algorithm for making sequence alignment but
for structures superimposed by other algorithm than Stamp and opened with
VMD.
I am using different software because for some proteins it is easier and
more convenient for me to use different than Stamp approach.
Thanks!

Best

tomek

On Mon, Nov 23, 2009 at 4:48 PM, Boris Steipe <boris.steipe_at_utoronto.ca>wrote:

> So you are looking for a sequence alignment based on a structure
> superposition?
>
> That's a bit of an ill-posed question. Imagine that your superposition
> places a residue in the middle between two residues of another structure?
> How would you define which one of them it is to be aligned with? If you
> think you can just choose whatever is the closest residue, what happens if
> more than one residue map onto one other? ...
>
> What you have in principle is a clustering problem where you need to ensure
> that each cluster contains exactly one and only one residue from each
> structure and the cluster metric is rmsd within the cluster.
>
> That said - since RMSD transformations are (usually) calculated from paired
> coordinate sets, it should be possible to get the information about which
> pairs were used from the original superposition algorithm. I.e. you would be
> better off not to try to recover the information from the result, but
> getting it from the procedure. Many superposition algorithms I have seen
> output that information somewhere.
>
> HTH,
> Boris
>
> PS: would the STAMP superposition even be significantly different from your
> original one?
>
>
>
>
>
> On 23-Nov-09, at 6:43 AM, Tomek Wlodarski wrote:
>
> Dear John,
>>
>> Thanks for reply!
>> Maybe I was not precise in my question...
>> Let say I have 10 structures opened with VMD, and they are already
>> structurally aligned by some other software. Now I am starting this Multiseq
>> plugin and I get multiseq window with sequences and I would like to have
>> those sequences aligned in such way that they resemble alignment of
>> structures. This alignment of sequences I can obtained after Stamp.
>> Can I obtained structure based sequence alignment without starting Stamp?
>> Just based on already superimposed structures?
>> Thanks!
>> Best
>>
>> tomek
>>
>> On Mon, Nov 16, 2009 at 10:28 PM, John Stone <johns_at_ks.uiuc.edu> wrote:
>>
>> Hi,
>> I'm puzzled by your question. I believe if if you do just
>> a sequence alignment, the alignment of the structures should have
>> no impact on the results. Even if STAMP realigned the structures,
>> the sequence alignment part of the result should be completely
>> independent of the structure alignment.
>>
>> Cheers,
>> John Stone
>> vmd_at_ks.uiuc.edu
>>
>> On Sat, Nov 14, 2009 at 03:02:43PM +0100, Tomek Wlodarski wrote:
>> > Dear all,
>> >
>> > I have structures structurally aligned by other program than STAMP.
>> > It is possible that when I will import all of the structures into
>> VMD,
>> > Multiseq plugin would generate sequence alignment based on how
>> structures
>> > are aligned?
>> > Thanks for any suggestions and ideas!
>> > Best!
>> >
>> > tomek
>>
>> --
>> NIH Resource for Macromolecular Modeling and Bioinformatics
>> Beckman Institute for Advanced Science and Technology
>> University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
>> Email: johns_at_ks.uiuc.edu Phone: 217-244-3349
>> WWW: http://www.ks.uiuc.edu/~johns/ <http://www.ks.uiuc.edu/%7Ejohns/>
>> Fax: 217-244-6078
>>
>>
>