Does anyone have a component for calculation the torsion angle in a small molecule ?

Hi,
I need to calculate torsion angles in a set of small molecules (not peptides).  I have looked into creating a component that would calculate the angle when given the 4 atom numbers.  PilotScript doesn't appear to have the vector geometry needed, and I haven't used the perl property calculator before...So, if anyone has one already, it would save me some time developing one.  (don't particularly want to redevelop the wheel here....)
Thanks !
Di
 
Answers 10
MH
MH 2016-12-02 Best Answer
Please give the attached protocol a try. Starting from the complex metal Atom it calculates the Torsion angles of all 5-membered rings where the metal Atom is part of.

Martin
comment imagecomment image
TD
TD 2016-11-18
Hi there -
Try the MolTorsion() function in PilotScript (and Java, Perl).
Can you share the component when you have it working?
Thanks!

 
DD
DD 2016-11-21
I'm having trouble because the MolTorsion() function needs to have the 4 atoms in order.  The middle 2 atoms being the ones that define the bond you're are measuring.  Unfortunately, the atom numbers in my molecules do not always go in the right order, and pipeline pilot functions seem to always present them to me in numerical order...I've tried several rather convoluted ways to get the correct order without much success....
MH
MH 2016-11-23
When you start with a .mol (or .sd) file, the Atom numbering in PP is following the order of Atoms in the file. This way you should be able to identify those 4 Atoms that you need for the calculation of the torsion angle. The attached protocol illustrates that.
comment imagecomment image
DD
DD 2016-11-23
In the attached file, I want the torsion angles of the bonds defined by atoms
a) 4,3,7,12
b) 17,16,20,25
and c) 29,28,32,37
When I use various methods to find these bonds (such as find the rings with the Ir in it, and delete the Ir atom) the resulting arrays are always ordered.  However, the torsion returned by MolTorsion(3,4,7,12) is not the same as the torsion returned by MolTorsion(4,3,7,12).
comment imagecomment image
MH
MH 2016-11-24
Deleting Atoms or bonds can cause atom/bond indices to Change. Maybe that is what is causing the confusion. It's not a trivial Task to identify the right Atoms to calculate a particular Torsion angle. If you have a defined functional Group, you could use a mapping to identify the set of Atoms. What are the criteria you want to apply to identify the right set of Atoms?
DD
DD 2016-11-28
Each of the angles I specified meet the following guidelines:
(1) It is part of a 5 membered ring that includes the iridium.
(2) If you take the iridium out of the set of atoms in that ring, the 4 remaining atoms define the torsion angle I'm interested in.  
However, the atoms usually are not numbered in order.  i.e. the first ring is (in pipeline pilot) 1,3,4,7,12 but to define the torsion angle correctly, you need to specify 4,3,7,12.
MH
MH 2016-11-28
We can write Molecular Toolkit code that walks through the 5-membered ring starting from the Ir and identifies the Atom numbers in the right order. It would of course be specific to that particular structure. So, this probably wouldn't help as it's not generic. We can only identify defined structural entities.
DD
DD 2016-11-28
Can you make the code generic enough to walk through any given ring and put the atoms in order ?
for instance if I give it a ring # ? or if I give it an array with the atom #s in it (i.e. I give it 1,3,4,7,12 and it give me 1,4,3,7,12 ) ? or if I give it a starting atom and say "give me the atoms of the rings that this atom is participating in, in order ? (i.e. for the iridium atom (atom # 1) it could give back an array like this:
1,4,3,7,12
1,17,16,20,25
1,29,28,32,37
)
DD
DD 2016-12-02
Hey !   It works !  Thank you !