What is Dreding Energy?

Hi Marvin users,





in my attempt to evaluate the ligand geometries predicted by 2 docking programs I used the term Dreiding energy, with the rational that the lower the energy, the better the stability of the conformation=>more physical. However I cannot find a detailed description about Dreiding energy and what attributes of the geometry it incorporates. Can anybody help me cite it in my report?





thanks in advance,


Tom

Hi Tom,





In the topic "Literature reference for dreiding" ( http://www.chemaxon.com/forum/viewpost14878.html ) you can find the article (S. L. Mayo, B. D. Olafson, W. A. Goddard III., J. Phys. Chem., 1990, 94, 8897-8909) which our Dreiding implementation based on.





I would like to note that in the Marvin version 5.1 we introduced major modifications in the Dreiding force field, so the energy values calculated with different Marvin versions are not comparable.





I also would like to note that any comparison based on a single energy calculation of structures which are not geometry optimized (using the actual force field) is likely also invalid. So if the geometries are generated using another tool then it is strongly recommended to optimize the structures before comparing energies.





If You have any further questions, dont hesitate to ask them.





regards,


Gabor

Hi Gabor,





thanks for the prompt response. I am a little bit confused about the thing you said concerning optimization. What I did is docked the same ligand onto the same receptor with 2 different programs, calculated the dreiding energy of the best pose predicted by each of them, after adding only polar hydrogens. Then I took the structure of the ligand from the DrugBank, added only polar hydrogens, did an energy minimization, calculated the Dreiding energy of the lowest energy conformer (very strict optimization limit), and compared this dreiding energy with those of the docking poses. If a docking pose differs a lot in its dreiding energy from the lowest energy conformer of the minimized structure, then it is less likely to be found in vivo.


As you might have noticed I didn't calculate the lowest energy conformer (I guess that's what you mean with the term "optimization") for the docking pose cause that would change the geometry and most of the stabilizing interaction between ligand - receptor will be lost.


Does my approach sound reasonable to you.





thanks,


Tom

Hi Tom,





Let me discuss the question of optimization first: the Dreiding force field (as the referred article describes) calculates the energy by summing energy components. These components are assigned to bond lengths, bond angles, dihedrals, etc. For example the bond angle component looks like E=1/2k(r-r0)^2 where r is the actual bond length found and r0 is the equilibrium bond length assigned by the force field and k is a constant value. As you can see a minute change in a bond length can have so much contribution to the calculated energy that it makes impossible to determine whether a conformation is chemically unfeasible or it is just a bit different from a local energy minimum. The optimization process takes a 3D initial structure and finds the nearest local energy minimum according to the force field.





Our "conformers" plugin can generate such low energy conformers from a 0D or 2D input structure, so you probably have to generate conformers at first for them. (The generated conformers are also optimized.) As far as I know, the DrugBank provides 2D structures. (I would like to note that a geometry optimization process currently can be invoked on a 2D input structure, but this not necessarily results in a valid low energy 3D structure.)





Since the force field does not consider the energy benefit of the protein-ligand interactions or any solvent dependent effect, the calculated energy does not necessarily reflect the validity of the docked conformer. I think a low energy found on a docked conformation can prove chemical feasibility but a high can not exclude it. Also, if a geometry optimization started from the docked conformation finds a local energy minimum near the starting structure can prove feasibility, but finding that the nearest local energy minimum is very different from the docked pose can not invalidate it. I would recommend you to check the docked conformations against their optimized conformers; do a conformational analysis on the ligand (generate multiple optimized conformers) and compare the generated conformers with the docked conformers.





To sum it up: if you find high energy calculated on a non-optimized docked conformation, then no conclusion can be made based solely on the energy calculations.





regards,


Gabor

tevang wrote:

Hi Marvin users, in my attempt to evaluate the ligand geometries predicted by 2 docking programs I used the term Dreiding energy, with the rational that the lower the energy, the better the stability of the conformation=>more physical. However I cannot find a detailed description about Dreiding energy and what attributes of the geometry it incorporates. Can anybody help me cite it in my report? thanks in advance, Tom

Hello

I just wanted to enquire the practical application of dreding energy in 2d qsar model. what does a positive value of DE implies.  Kindly help me as its a bit confusing to interpret results for me.

 

Thanks

Preeti

Hi Preeti,

The ChemAxon implementation of Dreiding forcefield is based on the original publication:

DREIDING: A Generic Force Field for Molecular Simulations Stephen L. Mayo, Barry D. Olafson, William A Goddard J. Phys.  Chem. 1990, 94, 8897-8909

although some modifications were made.

Our implementation uses bond length, bond angle, dihedral, inversion, van der Waals energy terms only. Many dihedral energy components were refined and extra dihedrals components are added. The original Dreiding forcefield is also extended so as to treat all possible atomtypes.

This forcefield energy similary to all other forcefields can be used to differentiate between different conformations of the same molecule in terms of stability. Lower the energy most stable the conformation is.  But beetween different molecules this energy is hard to interpret.

Please accept my apologies for the delayed answer.

Hope this helps.

Kind Regards,

Adrian