Bob,

Thank you for setting up an alignment feature for Jmol. The preliminary
alignment looks good. I need to try it with some other RNA structures
that are a little more complicated to see if it continues to give good
results. I hope to have time by Wednesday to try it. To answer your
questions:

Q: How important is that Jmol produce EXACTLY the result you show at
that site? -I would like to have a reasonable and believable alignment
of each RNA structure. The better the alignment, the more the users of
the site will be able to compare certain kink-turn RNA structures to
other kink-turn RNA structures. I am not looking to produce something
exactly as it is shown on my site, however it would be nice if they were
close and yield the same type of information.

Q: Do you have a reference to the algorithm that is being used by your
software? -I am using the Pair_Fit algorithm in PyMol. I do not have a
reference for it, but I will look to see if I can find it.  

Q: Would you like to work with me to make the ALIGN command work better
for your purposes? -I would like to work with you and make the ALIGN
command work in Jmol. I think it will not only benefit my website, but
might help other people out as well.

Again, thank you for your help and I look forward to working with you on
this Jmol feature. Have a good day.

Kersten



>>> Robert Hanson  15/03/10 1:06 PM >>>
Kersten, 

I have had some success in aligning your RNA fragments. Please get back
to me. 

Jmol 11.9.34 will align proteins, RNA, and DNA based on quaternion
frames using a new ALIGN command.

See files in

http://chemapps.stolaf.edu/jmol/docs/examples-11/align

for examples.

This is a little different from what you are doing there at your site,
though it is very close. Jmol is aligning the models to give the
smallest standard deviation for a set of quaternions representing the
orientations of the base pairs. As such, it does require prior knowledge
of which bases are to be matched. So the alignment command I used was
this:

load files "t1.pdb" "t2.pdb"
frame *
color group
set quaternionFrame "C"
qset1 = quaternion({41-46/1.1}) + quaternion({27-31/2.1}) 
qset2 = quaternion({77-82/2.1}) + quaternion({92-100/2.1})
align {2.1} {1.1} @qset2 @qset1

Where t1 and t2 are the starting files you have there for at
http://www.dundee.ac.uk/biocentre/nasg/kturn/sites/U4kturnSite.html

Let me explain what this is doing:

First, both files are loaded. 

The quaternionFrame "C" setting tells Jmol you want to align based on
nucleic acid bases. Another option is quaternionFrame "P", which aligns
based on the orientations of the tetrahedral phosphorus. But the
standard deviation was somewhat higher for that when I tried it.

qset1 and qset2 are each sets of quaternions. I only have preliminary
manuscripts on how these are set up, but the basic idea is that there is
a little xyz axis set (a "quaternion frame") on each base. You can
display those using:

select *
frame 1
draw quaternions
frame 2
draw quaternions
frame *

Jmol is optimizing the relative orientation of these quaternion frames. 

I have to do the concatenation business because the ordering of atoms in
the files are reversed for the two chains -- we need to match bases
77-82 and 92-100 in aligned_to_Kt7/3CC2_HmKt7.pdb (t2.pdb) with bases
41-46 and 27-31 in aligned_to_Kt7/1E7K_U4kt.pdb (t1.pdb), respectively. 

I could have left that out and just done

align {2.1} {1.1}

if these had instead been in sequential order.

The ALIGN command then aligns model 2 with model 1 by first finding the
optimum rotation to align the bases based on these quaternion sets, then
doing a simple center-of-geometry translation. It uses a very fast and
robust but rather new method referred to as the determination of the
"spherical mean" from:

Samuel R. Buss,Jay P. Fillmore:Spherical averages and applications to
spherical splines and interpolation.ACM Trans. Graph. 20(2): 95-126
(2001)


Of course, wh
ether or not this is an acceptable alignment is
questionable. Certainly it has never been done this way before. Still,
there are a lot of variables here, and I could imagine different ways of
going about it. It's not necessarily as simple as "find the smallest
root mean square deviation for the set of atoms" -- because of course we
have purines and pyrimidines mixed along the chains. So perhaps this is
an acceptable way to align the fragments. It's certainly very fast!

I would argue that Jmol's ALIGN command has given a better alignment of
the bases themselves; your alignment probably better matches exact atom
positions. Only you can decide which is "better."

Some questions:

Q: How important is that Jmol produce EXACTLY the result you show at
that site?

Q: Do you have a reference to the algorithm that is being used by your
software? 

Q: Would you like to work with me to make the ALIGN command work better
for your purposes?

The jar files needed for this analysis are for Jmol 11.9.34_dev, in 

http://chemapps.stolaf.edu/jmol/docs/examples-11/Jmol-11.zip

Hope to hear from you soon.

Bob


On Mon, Mar 8, 2010 at 9:43 AM, Kersten Schroeder  wrote:
Dear Robert,

I really enjoy Jmol and am currently using it on a website I help create
for the kink-turn RNA motif
(http://www.dundee.ac.uk/biocentre/nasg/kturn/kturns_all.php).

Do you plan to develop a way to align model 1 onto model 2 in Jmol
script? Pairfit in PyMol does it, so there may be a python code
available to help develop in Javascript. I want to align one model (of a
list of 30) to another model (of the same list of 30). Therefore,
pre-aligning elsewhere would need to be done about 900 times.

I look forward to your input. Thank you and have a good day.

Kersten Schroeder

Kersten Schroeder, Ph.D. 
MSI/WTB complex, 
Dow Street, 
University of Dundee, 
Dundee DD1 5EH 
UK 



    
             
************************************************************
Please       consider the environment. Do you really need to print this
email?    
  

------------------------------------------------------------------------------
Download IntelĀ® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Jmol-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jmol-users





-- 
Robert M. Hanson
Professor of Chemistry
St. Olaf College
1520 St. Olaf Ave.
Northfield, MN 55057
http://www.stolaf.edu/people/hansonr
phone: 507-786-3107


If nature does not answer first what we want,
it is better to take what answer we get. 

-- Josiah Willard Gibbs, Lecture XXX, Monday, February 5, 1900

************************************************************
Please consider the environment.  Do you really need to print this
email?
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Jmol-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jmol-users

Reply via email to